Cost per 1M Tokens
The standard unit for LLM API pricing: dollars per one million tokens, quoted separately for input and output.
Large language model APIs are priced per token, but a single token is a tiny fraction of a cent, so providers quote the rate per one million tokens (written 1M tokens, or sometimes per-MTok). Every price on this explorer uses that unit, in US dollars, and is split into at least an input rate and an output rate.
To turn a rate into a real cost, multiply the token count by the rate and divide by one million. A request with 40,000 input tokens on a $5.00/1M input model costs 40,000 ÷ 1,000,000 × $5.00 = $0.20 for input, plus the output tokens at the output rate. The token-cost calculator on the hub does this across every model for your own token volumes and monthly request count.
One caveat: a token is not a word. Roughly speaking, one token is about four characters of English or three-quarters of a word, but the exact count depends on each provider’s tokenizer. That means the same text can cost different amounts on two models with the same per-1M rate — see the tokenizer entry for why per-token price is not always an apples-to-apples comparison.
FAQ
How many words are in a million tokens?
Very roughly 750,000 words of English, since one token averages about three-quarters of a word. The exact figure depends on the model’s tokenizer.
How do I calculate the cost of a request?
Multiply your token count by the per-1M rate and divide by one million, separately for input and output, then add them. The hub calculator automates this across every model.