What a token is
In AI language models, a token is a small chunk of text that the model processes as a unit. Rather than reading whole words the way people do, models break text into tokens, which can be whole words, parts of words, or even single characters and punctuation, depending on the text. The model works with these tokens throughout, turning your input into tokens, processing them, and producing output tokens that become the response you see.
Why text is split into tokens
Breaking text into tokens gives models a consistent way to handle language, including rare words, misspellings, and many languages, by composing them from smaller pieces. A common word might be a single token, while an unusual word is split into several. This tokenization lets a model represent essentially any text with a manageable vocabulary of token pieces, which is more flexible and efficient than trying to have a unique entry for every possible word.
Tokens and cost
Tokens matter practically because AI services usually measure usage, and charge, by the number of tokens processed, counting both your input and the model's output. This means longer prompts and longer responses cost more, since they involve more tokens. Understanding that you pay per token, not per request, helps explain why concise prompts and controlling response length can meaningfully reduce the cost of using AI at scale.
Tokens and context limits
Tokens also define a model's context limit, the maximum amount of text it can consider at once, which is expressed as a number of tokens. If your input plus the expected output exceeds this limit, the model cannot handle it all, and older content may be dropped. Knowing that limits are measured in tokens helps you understand why very long documents or conversations can hit a ceiling and need to be shortened or split.
A rough sense of scale
While the exact mapping varies, a token often corresponds to roughly a short piece of a word, so a chunk of text contains more tokens than words. You do not need to count tokens precisely for everyday use, but having a rough sense, that longer text means more tokens, means more cost and more of the context limit used, helps you work with AI tools efficiently and avoid surprises with cost or length limits.
The bottom line
A token in AI is a small chunk of text, a word, part of a word, or a character, that language models process as a unit instead of reading whole words. Tokens matter because usage is usually priced per token and context limits are measured in tokens. Understanding them explains why prompt and response length affect both cost and what a model can handle at once.
Related on Skillo
See also: What is a context window? Explained, What is a large language model (LLM)?.
Sources
Published date reflects the original event date (2026-10-23). This article is original Skillo editorial written from the sources above; facts were verified in September 2026.
Leave a Reply