AI Token Price. Know what your tokens cost.

A token is a chunk of text, roughly three-quarters of an English word. Providers bill per million tokens, separately for the text you send (input) and the text the model writes (output). Different models split the same text into different numbers of tokens, so per-token prices are not perfectly comparable. See what one AI token costs today.

Each price comes from the provider's own pricing page, or for open-weight models sold by other hosts, the cheapest host listed on OpenRouter. Every model links to the page we checked, with the date. Every current price was checked between 13 Oct – 11 Nov 2026. We also compare our data with two public price feeds to catch changes; a difference only prompts us to re-check the provider's page and never changes a price by itself.

The model reads your input in parallel but writes output one token at a time, which uses far more compute per token. Output typically costs 3 to 8 times as much as input.

If you resend the same text (a system prompt, a document, the history of an agent session), most providers bill the repeated part at a much lower cache-read rate. Batch processing, where you accept results within hours instead of seconds, usually costs 50% less.

Your bill depends on how many tokens you send and receive, how much is served from cache and whether you use batch. Long prompts cost more on some models, and reasoning models bill their thinking as output tokens. The calculators estimate all of this.

Yes. The full dataset is published as JSON and CSV under CC BY 4.0. Credit AI Token Price with a link.