Public model catalog
Site-wide model catalog with USD list prices, no key required; any site may read it cross-origin, cached for 5 minutes.
/api/v1/models/pricingNo key required. Any site may read the response cross-origin, and it is cached for up to 5 minutes (Cache-Control: public, max-age=300).
Prices are USD list prices before the group multiplier; the price you actually pay for each model is in the signed-in model catalog. To see only which models a key can call, use GET /v1/models.
data[].platform is the protocol family the model is served under (for example openai, claude, gemini, kling), not the model vendor; see vendor for that. A model offered through several groups may appear under several platform values. For field details see Account and catalog API.
Returns
200Model catalog grouped by protocol family; with Access-Control-Allow-Origin: * and Cache-Control: public, max-age=300
0 on success
ok on success
Grouped by protocol family
The protocol family the model is served under, not the model vendor
Models in this protocol family
The model ID to pass
Display name
Model vendor, such as openai or google; may be absent
Series used by the console to group versions together; may be absent
Model category
Valueschatimagevideoaudioembedding
For example chat, reasoning, image_generation, image_edit, video_generation, tts
Tokens; not returned when unpublished
Unit for all prices in this entry: token per million tokens, second per second of video, image per image, character per million billed characters
Valuestokensecondimagecharacter
Input price, USD / price_unit; speech models use only input
Cached input price, USD / price_unit; absent when caching does not apply
Output price, USD / price_unit
Present when there is a long-context tier: when input tokens exceed threshold, the whole request is billed with the three multipliers
Input token threshold
Input multiplier
Cached input multiplier
Output multiplier
Per-image price by resolution tier, for example {"1k": …, "2k": …, "4k": …}
Video prices tiered by resolution, audio on/off and reference media on/off. When price_unit is token they are per million video tokens; when second, per second. The key names are kept for historical reasons