Context is the text the model can see

A context window is the amount of text, and sometimes images, sent with one request. The model does not remember your last meeting unless that text is in the request or in a store you explicitly search.

Bigger is not automatically better. Extra text costs money, adds latency, and can bury the instruction you care about. Put the task and the constraints first. Paste the source the model must use, not a dump of everything you have.

If an answer ignores a fact you thought you supplied, check whether that fact was in the request before you blame the model.