What is the context window?
The context window is the amount of text a model can hold in view for a single answer, and once the input goes past that limit, whatever falls outside it is simply not seen.
Every time a model answers, it's working with a fixed-size window of text: the instructions behind the scenes, the conversation so far, any material that's been pulled in to help answer the question, and the question itself. That window has a limit, measured in small chunks of text called tokens. Content that doesn't fit gets trimmed or dropped rather than remembered anyway, the model genuinely does not see it.
This has a practical consequence for anyone trying to get a page read accurately by an AI tool: a long page competes for limited room. If the useful part, the price, the answer, the key fact, sits far down the page, it's more likely to be truncated or crowded out than if it appears early. It's also why a long conversation can seem to make an AI 'forget' something said much earlier: that detail hasn't vanished from the world, it's just fallen outside the window for that answer.
Related
- TokensTokens are the small chunks, often pieces of words rather than whole words, that a language model actually reads and produces, and they're the unit that usage limits, pricing, and the context window are all measured in.
- GroundingGrounding is when a model answers using documents retrieved at the time of the question, rather than only from what it absorbed during training.
- Retrieval-augmented generation (RAG)Retrieval-augmented generation is the technique of retrieving relevant documents and giving them to a model as context before it answers.
- PromptA prompt is the full text handed to a model right before it responds, and for any real AI product that's almost always more than just what the user typed.
Want to know where you actually stand on this? Run a free visibility check or try the free tools.