How much the model can hold in mind at once.
The context window is the maximum number of tokens a model can consider in a single request — the prompt, any attached documents, the conversation so far and the answer being written, all counted together.
Windows have grown from a few thousand tokens to a million and more, which is what makes whole-codebase and whole-contract analysis possible.
Long context is not free: attention cost grows faster than linearly with length, so a huge prompt is slower and dearer, and models often attend less reliably to the middle of very long inputs.