Token
Also: tokens, token usage
A token is the basic unit Claude uses to read and write text. On current models the ratio is a bit over half a word per token — Anthropic changed the tokenizer with Claude Opus 4.7, so a million tokens holds roughly 555,000 words. On models older than that, the ratio is about three-quarters of a word per token. Everything you send to Claude (your message, system prompt, uploaded documents) and everything Claude sends back counts as tokens. Tokens determine both your usage limits and your API costs.
In practice
You ask Claude a question that's 50 words. Claude responds with 200 words. That exchange is roughly 450 tokens on a current model. Your API bill reflects that usage. Long conversations, big documents and verbose system prompts all add tokens — and tokens are directly what you pay for. When the estimate is driving a budget rather than a guess, use the count_tokens endpoint instead of a ratio.
Related concepts
Where Token shows up
2 articlesTokens are what you pay for. Here are the practical things you can do to use fewer of them — from how you prompt to which model you choose.
Tokens are how language models read and write text — and how every AI API charges you. Understanding them turns abstract pricing into something you can predict and control.