Hi models.dev maintainers,
Thank you for maintaining a public model metadata dataset — it's valuable infrastructure for routing, cost estimation, and developer tooling.
While cross-auditing model metadata against primary provider sources, we found the mistral-medium-latest entry appears to diverge from Mistral's current published prices (https://mistral.ai/pricing/api/, accessed 2026-07-04):
- models.dev:
cost.input: 0.4, cost.output: 2
- Mistral's pricing page: "Mistral Medium 3.5 ... Input (/M tokens) $1.5 Output (/M tokens) $7.5 mistral-medium-latest"
Context: we ran this audit while building FactQuire (https://factquire.com), a source-verified facts feed for commercial LLM APIs where every price/limit entry carries a verbatim quote from the primary source (machine-readable feed: https://factquire.com/feed.json). The audit cut both ways — it also caught an error in our own data (gemini-2.5-flash context window), which we corrected publicly.
Happy to open focused PRs with source-backed corrections mapped to the models.dev schema if that's useful.
Hi models.dev maintainers,
Thank you for maintaining a public model metadata dataset — it's valuable infrastructure for routing, cost estimation, and developer tooling.
While cross-auditing model metadata against primary provider sources, we found the
mistral-medium-latestentry appears to diverge from Mistral's current published prices (https://mistral.ai/pricing/api/, accessed 2026-07-04):cost.input: 0.4,cost.output: 2Context: we ran this audit while building FactQuire (https://factquire.com), a source-verified facts feed for commercial LLM APIs where every price/limit entry carries a verbatim quote from the primary source (machine-readable feed: https://factquire.com/feed.json). The audit cut both ways — it also caught an error in our own data (gemini-2.5-flash context window), which we corrected publicly.
Happy to open focused PRs with source-backed corrections mapped to the models.dev schema if that's useful.