kUsageModelPrices top-level constant

ModelPriceTable const kUsageModelPrices

The station's declared prices — read ONLY when a harness reports tokens but no billed cost.

gpt-5.6-sol carries the OpenAI list rates verified 2026-09-04 (developers.openai.com/api/docs/models/gpt-5.6-sol): input $4.00, cached input $0.40, output $20.00 per million; cache CREATION is $5.00 because that page bills a cache write at 1.25× input.

The claude models are deliberately ABSENT: their envelope reports total_cost_usd, which stays authoritative — a reported price is never second-guessed by a table that can drift.

Implementation

const ModelPriceTable kUsageModelPrices = <String, ModelTokenPrice>{
  'gpt-5.6-sol': ModelTokenPrice(
    inputUsdPerMillion: 4,
    cacheReadUsdPerMillion: 0.4,
    cacheCreationUsdPerMillion: 5,
    outputUsdPerMillion: 20,
  ),
};