Token
The small chunk of text a language model actually reads and predicts — often a whole word, a piece of a word, or a punctuation mark. Models measure and price everything in tokens, not characters.
Think of It Like This
Tokens are the LEGO bricks of text — models build and bill in bricks, not whole sentences.
A model never sees raw characters. A tokenizer first splits your text into tokens drawn from a fixed vocabulary, and each token is then mapped to an embedding the model can do math on. A handy rule of thumb: one token is roughly four characters of English, so a page of prose is about 500 tokens. That count is what sets both how much you can fit in the context window and the bill you pay per call.