Refresh OpenAI pricing and support context/service-tier rates #86

Summary

Gateway cost estimation uses stale and incomplete OpenAI pricing metadata. New GPT-5.6 models have no pricing, while several GPT-5.5/GPT-5.4 constants do not match the current Standard short-context table and appear to mix Batch or long-context rates.

Live evidence

Observed through the running gateway_usage_snapshot:

• total tokens: 176,092,215;
• total requests: 1,452;
• requests missing model pricing: 1,448;
• estimated cost: $0.047209875;
• recent 500-event sample: 497 requests used openai/gpt-5.6-sol, whose runtime pricing metadata is null.

Runtime metadata currently contains prices for gpt-5.4, gpt-5.4-mini, and gpt-5.5, but no prices for gpt-5.6-sol, gpt-5.6-terra, or gpt-5.6-luna.

Root cause

builtin_model_metadata() and OFFICIAL_FAST_PRICING are static. ModelPricing only contains one input, cached-input, and output rate per model. Current OpenAI pricing also varies by short/long context and Standard/Batch/Flex/Priority processing, with cache-write pricing for applicable models.

Expected behavior

• Pricing matches the current official OpenAI source.
• Estimation selects the rate for the actual service tier and context band.
• Unknown tiers or missing data remain unpriced.
• Source URL and retrieval/version timestamp are visible in metadata.

Acceptance criteria

• Add current pricing for gpt-5.6-sol, gpt-5.6-terra, and gpt-5.6-luna.
• Reconcile GPT-5.5, GPT-5.4, and GPT-5.4-mini with current Standard pricing.
• Represent short/long-context rates where applicable.
• Represent Standard, Batch, Flex, and Priority separately, or limit estimation to the observed tier.
• Record enough telemetry to select the correct tier; fail to unknown when unavailable.
• Do not treat Fast/Priority aliases as Standard traffic.
• Add table-driven tests for aliases, context threshold, cache input, output, and tiers.
• Add a documented update/check procedure for newly discovered official models.

Primary references

• https://developers.openai.com/api/docs/pricing
• https://developers.openai.com/api/docs/guides/latest-model.md

Relevant code

• src-tauri/src/models.rs
• src-tauri/src/gateway/telemetry.rs