All termsGlossary

What is token?

A token is the small unit of text that an AI language model reads and writes. It can be a whole word, part of a word or a punctuation mark.

Why it matters

Tokens are how language models are limited and charged. The context window is counted in tokens, and usage is normally billed by the number of tokens read and written. Reply speed depends on them too, because a model produces its answer one token at a time. A team that understands tokens can estimate cost and keep replies quick.

How it works

Before text reaches the model, a tokeniser splits it into pieces from a fixed vocabulary. Common short words are usually one token each, while long or unusual words are split into several. The model works with these pieces, not with letters or whole sentences. Token counts differ between languages and between models, so the same message can use more tokens in one language than another.

Example

A customer writes, “I can’t log in.” The tokeniser might split this into six tokens, with the contraction and the full stop counted separately. The agent’s instructions add 800 tokens and two help articles add 1,200. The request is therefore about 2,000 tokens before the model writes a word. Trimming the articles to the relevant paragraphs brings it down to about 1,400.

Get Started

Let’s build AI agents, together.

Talk with us to see the top three use cases for AI agents at your business, and how quickly you could go live.

Live in 7 days · Works inside the tools your team already uses · Hands off with full context