Context window
The amount of information a model can handle in one request, usually measured in tokens. How inputs and generated output share that limit depends on the model and service.
For example
A long conversation and its attached documents can use much of a model's context window before it begins answering.
By Dave AI Editorial · Prompts & context