What is a Context Window?
A context window is the maximum amount of information, measured in tokens, that a model can process within one interaction or request. It represents the strict upper limit on how much text the AI can "see" at any given moment.
How does it work?
When you chat with an AI, the system sends your new prompt along with the history of the conversation to the model. If a model has a context window of 100,000 tokens, and your conversation reaches 100,001 tokens, the system must drop the oldest messages to fit the new prompt in.
What is it commonly confused with?
Do not describe the context window as "permanent memory." The model itself does not remember you between sessions. When you start a new chat, the context window is entirely empty. While some AI applications feature "memory systems" that save facts about you to a database, that is an application-level feature, completely distinct from the model's actual context window.
Why does it matter?
Larger context windows allow users to upload massive codebases, entire books, or dozens of financial reports simultaneously for the AI to analyze in a single pass without losing the thread of the conversation.