← All terms
FundamentalsLast reviewed
Token
A token is the smallest unit of text a language model works with — usually a word or part of a word. The length and cost of a request are measured in tokens.
How it works
Language models do not read text letter by letter but in tokens. A token can be a whole word, a part of a word or a punctuation mark. As a rule of thumb, one token in English is about four characters. German texts often need more tokens because of long compound words.
A practical example
A ten-page document consists of several thousand tokens. If you send it to a model and ask for a summary, input and answer together count — for the price and for the context window.
What you should know
- When you use a model through an API, billing is usually per token, separately for input and output.
- Short, clear instructions save tokens, time and money.
- How many tokens a text has depends on the model. Providers offer counting tools for this.
Matching tools