What is a context window in LLMs and why does token limit matter?

The context window defines the short-term working memory of a large language model during a conversation or task run.

Why Context Windows Matter

If a prompt exceeds the model's context capacity (e.g. 128k or 200k tokens), older tokens are truncated, resulting in forgotten instructions or loss of conversation state.

Compare context window capacities across frontier models in our AI Model Comparison Tracker.