In plain English
A token is the basic unit of text a language model processes. It can be a whole word, part of a word, or a punctuation mark, and models convert your text into tokens before working with it. Models have a limit on how many tokens they can handle at once, and usage is usually billed per token.
Why it matters
Tokens explain two practical things at once: why there is a limit on how much a model can read in one go, and why longer prompts and answers cost more.
A worked example
A common word like 'the' is usually one token, while a long or unusual word may be split into several, so a 1,000-word document is roughly 1,300 tokens.
Common confusion
A token is not exactly a word. Because words can split into several tokens, token counts and word counts do not line up neatly.