Guide
Glossary
Token
A token is the small unit of text an AI model reads and writes, usually a whole word, part of a word, a number or a punctuation mark. Models split text into tokens before processing it. As a rule of thumb from OpenAI, 100 tokens is about 75 English words, though figures, tables, code and other languages often take more. Tokens matter because they measure most limits and costs: how much a model can read at once (its context window), how long its answer can be, and the cost of using a model through its API, which is usually priced per million tokens.
Also called AI token, LLM token
Last updated:
Example
In an accounting firm
An accounting firm wants an AI tool to review a 60-page set of client financial statements. At about 500 words a page, that is roughly 30,000 words, or around 40,000 tokens by the usual rule of thumb, and more because the pages are dense with figures. Knowing this helps the firm choose a tool whose context window fits the whole document.
Guides that explain it in context
Secure by design. Set up correctly. Fully managed.
Talk to us before you commit to anything
Start with a free 45-minute discovery call. We look at your systems and priorities, then recommend a first step with a fixed scope, or tell you if we are not the right fit.