Context Window
Also known as: context length, token window.
In plain English
Tokens are roughly 3-4 characters or 0.75 of an English word. A 1M-token context window holds approximately 750,000 words — about 1,500 pages of text. The window has to fit your prompt + the model's response, so the practical input limit is slightly smaller than the headline number. Longer windows enable use cases that were previously impossible: feeding entire codebases, books, or contracts to the model in one prompt.
Example
An analyst feeds a 280-page contract to Gemini 3 Pro (2M context). The model holds the entire document and answers questions like 'which clauses conflict with our standard MSA template?' without any document chunking. The same prompt would have to be split across multiple calls on GPT-5.2's ~400k window.
Context Window in Vincony
Vincony exposes the long-context capabilities of every model in its catalog. Use Compare Chat to send a very long document to multiple models at once and pick the best response.
Compare context windowsTry it — 750+ distinct models across 80+ providers on one account
Vincony bundles GPT-5, Claude, Gemini, Perplexity Sonar Pro, DeepSeek, Mistral, and 750+ other models on one $0/month account.