[BOT ISSUE] Databricks: add missing databricks-gpt-5-5-pro, databricks-gpt-5-4-mini, databricks-gemini-3-5-flash

[github-actions]

Gap

Type: Missing models

Three Databricks Foundation Model API pay-per-token endpoints are listed on official Databricks documentation but absent from packages/proxy/schema/model_list.json. The base models already exist in the catalog for other providers, but the Databricks variants have no entries.

Missing Models

Model ID	Display Name	Base Model	Context Window	Max Output
databricks-gpt-5-5-pro	GPT-5.5 Pro	gpt-5.5-pro (line 90)	1,050,000	128,000
databricks-gpt-5-4-mini	GPT-5.4 mini	gpt-5.4-mini (line 157)	272,000	128,000
databricks-gemini-3-5-flash	Gemini 3.5 Flash	gemini-3.5-flash (line 10920)	1,048,576	65,536

Context

• The catalog already has 36+ databricks-* entries including databricks-gpt-5-5, databricks-gpt-5-4, databricks-gpt-5-4-nano, databricks-gemini-3-1-flash-lite, databricks-gemini-3-1-pro, etc.
• The base models exist under other providers: gpt-5.5-pro (openai, azure), gpt-5.4-mini (openai, azure), gemini-3.5-flash (google, vertex)
• Existing open issue [BOT ISSUE] Databricks: add missing GPT Codex models (5.3, 5.2, 5.1 variants) #760 covers Codex variants; this issue covers three different model families

Suggested Catalog Entries

"databricks-gpt-5-5-pro": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "displayName": "GPT-5.5 Pro",
  "reasoning": true,
  "max_input_tokens": 1050000,
  "max_output_tokens": 128000,
  "available_providers": [
    "databricks"
  ]
},
"databricks-gpt-5-4-mini": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "displayName": "GPT-5.4 mini",
  "reasoning": true,
  "max_input_tokens": 272000,
  "max_output_tokens": 128000,
  "available_providers": [
    "databricks"
  ]
},
"databricks-gemini-3-5-flash": {
  "format": "openai",
  "flavor": "chat",
  "multimodal": true,
  "displayName": "Gemini 3.5 Flash",
  "reasoning": true,
  "reasoning_budget": true,
  "max_input_tokens": 1048576,
  "max_output_tokens": 65536,
  "available_providers": [
    "databricks"
  ]
}

Token limits are based on the base model entries. Pricing omitted per Databricks provider constraints (no stable public per-model pricing source).

Verification Checklist

Check	Status	Detail
Cross-source	✅	All three model IDs confirmed on Databricks Foundation Model Overview (pay-per-token regional table) AND databricks-gpt-5-5-pro independently confirmed in Databricks Responses API docs
Already fixed	✅	All three IDs absent from model_list.json — confirmed by grep
ID format	✅	Follows databricks-{model} naming convention used by 36+ existing entries (e.g., databricks-gpt-5-5, databricks-gemini-3-1-flash-lite)
Duplicate check	✅	No existing open issue covers these three model IDs (issue #760 covers Codex variants only)

Verification Notes

Field	Source	Notes
Model IDs	Databricks Foundation Model Overview	Listed in pay-per-token endpoint regional availability tables
format: "openai"	Catalog convention	All Databricks models use openai format
flavor: "chat"	Catalog convention + Databricks docs	Chat/completion models
reasoning: true	Base model entries	gpt-5.5-pro, gpt-5.4-mini, gemini-3.5-flash all have reasoning: true
reasoning_budget: true	Base gemini-3.5-flash entry	Only Gemini 3.5 Flash base model has this field
multimodal: true	Base model entries	All three base models have multimodal: true
Token limits	Base model entries	Matched to each base model's token limits
Pricing	Not published	Databricks does not publish stable per-model token pricing

Local Files Inspected

• packages/proxy/schema/model_list.json — confirmed all three model IDs absent; base models at lines 90, 157, 10920; existing databricks-* entries at lines 12142–12376

{
  "kind": "missing_model",
  "provider": "databricks",
  "models": ["databricks-gpt-5-5-pro", "databricks-gpt-5-4-mini", "databricks-gemini-3-5-flash"],
  "status": "active",
  "model_specs": {
    "databricks-gpt-5-5-pro": {
      "format": "openai",
      "flavor": "chat",
      "multimodal": true,
      "displayName": "GPT-5.5 Pro",
      "reasoning": true,
      "max_input_tokens": 1050000,
      "max_output_tokens": 128000,
      "available_providers": ["databricks"]
    },
    "databricks-gpt-5-4-mini": {
      "format": "openai",
      "flavor": "chat",
      "multimodal": true,
      "displayName": "GPT-5.4 mini",
      "reasoning": true,
      "max_input_tokens": 272000,
      "max_output_tokens": 128000,
      "available_providers": ["databricks"]
    },
    "databricks-gemini-3-5-flash": {
      "format": "openai",
      "flavor": "chat",
      "multimodal": true,
      "displayName": "Gemini 3.5 Flash",
      "reasoning": true,
      "reasoning_budget": true,
      "max_input_tokens": 1048576,
      "max_output_tokens": 65536,
      "available_providers": ["databricks"]
    }
  },
  "source_urls": [
    "https://docs.databricks.com/aws/en/machine-learning/model-serving/foundation-model-overview",
    "https://docs.databricks.com/aws/en/machine-learning/model-serving/query-openai-responses"
  ]
}

Summary: Three Databricks pay-per-token models are missing from the catalog. The base models exist under other providers; these Databricks-specific entries need to be added following the existing databricks-* naming convention.

[github-actions]

Prepared catalog updates for databricks-gpt-5-5-pro, databricks-gpt-5-4-mini, databricks-gemini-3-5-flash. Added models: databricks-gpt-5-5-pro, databricks-gpt-5-4-mini, databricks-gemini-3-5-flash.

Verification sources: https://docs.databricks.com/aws/en/machine-learning/model-serving/foundation-model-overview, https://docs.databricks.com/aws/en/machine-learning/model-serving/query-openai-responses
Verified metadata applied:

• databricks-gpt-5-5-pro: format=openai, flavor=chat, displayName=GPT-5.5 Pro, max_input_tokens=1050000, max_output_tokens=128000, pricing=not provided, providers=databricks, reasoning=true, multimodal=true
• databricks-gpt-5-4-mini: format=openai, flavor=chat, displayName=GPT-5.4 mini, max_input_tokens=272000, max_output_tokens=128000, pricing=not provided, providers=databricks, reasoning=true, multimodal=true
• databricks-gemini-3-5-flash: format=openai, flavor=chat, displayName=Gemini 3.5 Flash, max_input_tokens=1048576, max_output_tokens=65536, pricing=not provided, providers=databricks, reasoning=true, reasoning_budget=true, multimodal=true

Included in the daily batch PR: #989