Context Window

A context window is the maximum amount of text (measured in tokens) that an AI model can process in a single inference call. Modern frontier models support 200K–1M tokens, but effective recall and reasoning quality often degrade well before that limit.