What are tokens?
Tokens are the small chunks, often pieces of words rather than whole words, that a language model actually reads and produces, and they're the unit that usage limits, pricing, and the context window are all measured in.
Before a model processes any text, that text gets broken into tokens: a token might be a short whole word, part of a longer word, a punctuation mark, or a space. It's not a one-token-per-word split, so token counts and word counts diverge, especially with unusual words, other languages, or heavy formatting.
This matters mainly because of what gets measured in tokens: how much text fits in the context window, how AI products price API access, and how long a response is allowed to run. None of that is a reason to write for token counts the way people once wrote for keyword density, but it does mean a page's actual length has a literal cost: the longer and more padded a page is, the more of that limited budget it eats before a model even reaches the part worth citing.
Related
- Context windowThe context window is the amount of text a model can hold in view for a single answer, and once the input goes past that limit, whatever falls outside it is simply not seen.
- InferenceInference is the moment a model actually generates an answer to one specific question, separate from training, which is the earlier, one-time process of building the model in the first place.
- Large language modelA large language model is a program trained on huge amounts of text to predict what word comes next, and that single skill is what lets it write, summarize, and answer questions.
Want to know where you actually stand on this? Run a free visibility check or try the free tools.