Back to the AI glossary Models
Tokens & context windows
The unit of measurement for text in AI models and their processing limit.
Explanation
Tokens are the smallest units of text that an LLM processes – approximately ¾ of a word. The context window defines how many tokens the model can "see" at the same time.
How it works
Each text is broken down into tokens. A model with 128K context windows can process around 100,000 words at the same time - that's about the equivalent of an entire book.
Example
If you're analyzing a 50-page document, it has to fit in the context window. If it is too large, it must be divided (→ chunking).
Why it matters
The context window determines how much information a model can take into account at the same time - and influences costs and quality.