Context Window
The maximum number of tokens — prompt plus response — a model can consider at once, ranging from 200K to over 1M.
The context window is the total number of tokens a model can hold in view for a single request: your entire prompt, the conversation history, any retrieved documents, and the response it generates all have to fit inside it. Models in this explorer range from 200K-token windows up to more than one million tokens.
A larger window lets you pass more source material — whole codebases, long documents, extended chat histories — without truncating or chunking. But window size is a capacity limit, not a price: filling a 1M-token window still bills every one of those tokens at the input rate, so a big context is a convenience that can get expensive fast. Some providers add a long-context surcharge above a threshold, which is a separate pricing concept.
This explorer lists the context window per model where the provider publishes it; a dash means it was not confirmed on the official pricing page and we did not guess. Note that maximum output tokens is a separate, usually smaller limit — the window bounds the whole request, while the output cap bounds just the response. Check both when a task needs either a very long input or a very long answer.
FAQ
Does a bigger context window cost more?
The window itself is not a price, but filling it costs more because every token in the prompt is billed. Some models also add a long-context surcharge above a token threshold.
Is the context window the same as max output tokens?
No. The context window bounds the whole request (prompt plus response); max output tokens caps only the generated response, and is usually much smaller.