Text
Priced per 1M input and output tokens. Cache hits are billed at a lower input rate where the provider supports them.
Model catalogue
Every model, one key.
Search by name or model id, filter by provider group or modality, and sort by price. Text prices are per 1M tokens; media models bill per image or per generated second.
Prices are provider reference values, last checked 22 Jul 2026. See full pricing for cache-hit rates, long-context tiers and per-second media billing.
Reading the table
Billing units differ by modality.
Image
Priced per generated image, sometimes with a pixel threshold above which a higher rate applies.
Video
Priced per generated second, with separate rates for 720P and 1080P output.
Audio
Priced per 1M audio tokens, with text tokens in the same request billed separately at the text rate.