Context window

The maximum amount of text, measured in tokens, that a model can consider at once.

concept

The context window is the model’s working space for a single request. Your prompt, any documents or conversation history sent with it, and the model’s reply all have to fit inside it.

When a conversation grows longer than the window, something has to give: older parts may be dropped or summarised, and the model can no longer take them into account.

A larger window lets a model work with longer material in one go, but it does not guarantee that every part gets equal attention.

Example

After a very long chat, a question about the first message may get a vague answer, because that message has slipped out of the window.

Related