Pricing
Pay as you go. You pay each model's per-token rate (set by its providers) plus a small platform fee — no seats, no minimums, no monthly commitment.
Pass-through token pricing
Every model lists its input / output price per 1M tokens. We route to the cheapest healthy provider, so you pay the live market rate — not a marked-up flat rate.
View model pricesPrepaid credits
Top up credits and draw down as you use the API. A 5% fee applies when purchasing credits; usage itself is billed at pass-through rates with no extra per-request charge.
Add creditsCaching & routing
Prompt caching can cut effective input cost by 60–80% on repeated context. Use :floor to always take the cheapest provider, or set a hard max_price per request.
Prices shown across the site are per 1M tokens in USD and update as providers change their rates. Enterprise volume discounts are available — contact sales.