Context Window Visualizer
See how much of a model’s context window your prompt actually uses, and which part of the conversation is consuming it.
Context utilisation
160,000 / 1,000,000
16.0%
Headroom
840,000 tokens free. Remember the model's reply comes out of the same budget — leave room for it.
Context is a ceiling, not a promise
Retrieval accuracy typically falls off well before the advertised maximum. A model quoting a 1M window will not reliably recall a detail buried at 800K.
Frequently asked questions
What does the Context Window Visualizer show?
You enter token counts for the system prompt, user messages and assistant messages, choose a model, and the tool shows how much of that model’s context window each part uses and how much is left free. It warns at 80% and again at 95%, and reports overflow if you exceed the window.
Where do I get the token counts to enter?
This tool takes token counts, not raw text. Use the Token Counter to measure each part of your prompt, then enter the numbers here. Remember that Claude and Gemini counts from the Token Counter are estimates.
Does the context window include the model’s reply?
Generally yes. The window covers the prompt and the generated response together, so you need to leave headroom for the output you expect. A prompt that fills the window leaves little or no room for an answer.
Why leave space even before the window is full?
The 80% and 95% markers are warning levels, not hard limits. Models tend to recall details less reliably as the context grows, so staying well below the limit is safer than treating the full window as usable.