Models & pricing

The live catalog with per-model prices is on the Connect page in your cabinet (and via GET /v1/models).

• Availability depends on your group — you see only the models your account can call.
• Prices are per 1M units (input / output / cache read / cache write), shown in USD — the amount you actually pay.
• Cache reads are billed at a fraction of input; writes (where a provider charges them) are listed separately.
• Protocols: each model row shows which protocols it supports (OpenAI / Anthropic).

A fuller pricing reference with examples will live here.