fetching-models-pricing

name	fetching-models-pricing
description	Refresh per-model pricing tables for Floway providers whose upstream does not bill per token or publish usable token rates, especially Copilot, Codex, Claude Code, and Ollama. Manual research procedure; no script.

Fetching Models Pricing

Maintain the notional per-token rate cards in:

Provider	Table	Live catalog	Preferred rate source
Copilot	packages/provider-copilot/src/pricing.ts	Copilot /models	model vendor's first-party API
Codex	packages/provider-codex/src/pricing.ts	authenticated /codex/models	OpenAI API pricing
Claude Code	packages/provider-claude-code/src/pricing.ts	authenticated Anthropic /v1/models	Anthropic API pricing
Ollama	packages/provider-ollama/src/pricing.ts	/api/tags + /api/show	vendor API or a credible commodity host

These providers are subscription-backed or self-hosted. Floway records notional API-equivalent value so the usage dashboard remains comparable.

Procedure

1. Fetch the provider's live catalog and diff its ids against the table's string and RegExp keys. Record new, retired, and renamed models.

2. Find a defensible rate source for every new id:

   • Prefer the model vendor's first-party API.
   • For open weights with no vendor API, use the cheapest credible commodity host that publishes the required dimensions.
   • For retired versions, use a permalink or dated archive from when that version was current.

3. Cross-check at least two sources. models.dev remains useful as an independent comparison under its external cost field:

   curl -s https://models.dev/api.json | jq '.<provider>.models["<id>"].cost'
   

   OpenRouter prices below first-party rates are usually mirror-host prices, not the canonical vendor rate.

4. Author one ModelPricing with modelPricing and pricingEntry:

   modelPricing(
     pricingEntry({ input: 2.5, input_cache_read: 0.25, output: 15 }),
     pricingEntry(
       { input: 5, input_cache_read: 0.5, output: 22.5 },
       { inputTokens: { operator: 'gt', value: 272000 } },
     ),
   )
   

   Every entry is one exact selector coordinate plus explicit USD-per-million- token rates. Follow these invariants:

   • Declare exactly one Base entry without a selector.
   • Give every entry the same rate dimensions as Base.
   • Never merge entries or inherit individual cache/image rates from another dimension. A dimension absent from Base is unpriced everywhere.
   • Treat serviceTier as an open-string equality coordinate.
   • Treat inputTokens gt / gte thresholds as whole-request bands, not marginal token buckets.
   • Threshold-only entries are global. Thresholds combined with equality coordinates apply only inside that exact scope. Runtime selects the highest matching global or scoped threshold, then performs one exact rate lookup.
   • A missing exact selector uses the whole Base vector. Do not synthesize an undocumented Cartesian combination.
   • Return null when no defensible price exists. Never extrapolate from an adjacent version or similarly named model.

5. Increment MODEL_CATALOG_REVISION in packages/gateway/src/data-plane/providers/models-cache.ts. Static pricing is serialized inside cached ProviderModel rows; a mismatch makes every older row cold before TTL evaluation.

6. Add boundary tests for exact ids, aliases, dated releases, RegExp coverage, threshold edges, and Base fallback through priceRequest.

7. Run all affected provider tests, typecheck, lint, and the full test suite.

8. If an existing rate changed, use backfill-model-pricing for the intended historical usage slice. Catalog revisioning changes future snapshots; it does not rewrite recorded unit prices.

Catalog Revision Policy

MODEL_CATALOG_REVISION versions the complete persisted ProviderModel contract, not only pricing tables. Increment it for any code change that alters code-derived catalog metadata or its serialized representation. Upstream-only catalog changes do not require a bump because normal TTL refreshes fetch them.

The revision is global by design. A mismatch blocks on a fresh fetch and never falls back to the incompatible stored row. Successful fetches overwrite the row with the current revision; failed fetches leave the old row present but ineligible.

Provider Identity

• Copilot usage stores raw variant suffixes such as -high, -xhigh, and -1m in model_key; its pricing lookup normalizes them to the public id.
• Claude Code resolves pricing from the dated raw upstream id before catalog aliases are merged into public ids.
• Codex and Ollama use the raw upstream slug directly.

Source Cautions

• Ignore LiteLLM's zero-valued ollama/* entries; those are placeholders, not market prices.
• Do not confuse Ollama library labels such as Light, Medium, High, or Extra High with token prices; they are subscription GPU-time weights.
• Do not use a cheaper OpenRouter mirror when the vendor itself sells the model; that is another host's price.
• Verify ambiguous version names against release notes before sharing a rate.
• Keep a permalink or stable official URL beside every vendor constant and document non-obvious source choices next to the table entry.