CONTEXT WINDOW COMPARISON
How much text fits in each AI model.
Every model's context window, sorted largest first, with token limits translated into approximate English words and pages. Click any model to open its token counter and check whether your actual text fits. Data refreshed 2026-08-11.
| Model | Provider | Context window (tokens) | ~Words | ~Pages |
|---|---|---|---|---|
| Llama 4 Scout | Meta | 10,000,000 | ~7,500,000 | ~15,000 |
| Gemini 3.6 Flash | 1,048,576 | ~786,000 | ~1,573 | |
| Gemini 3.5 Flash-Lite | 1,048,576 | ~786,000 | ~1,573 | |
| Gemini 3.5 Flash | 1,048,576 | ~786,000 | ~1,573 | |
| Gemini 3.1 Pro Preview | 1,048,576 | ~786,000 | ~1,573 | |
| Gemini 3.1 Flash-Lite | 1,048,576 | ~786,000 | ~1,573 | |
| Gemini 2.5 Pro | 1,048,576 | ~786,000 | ~1,573 | |
| Gemini 2.5 Flash | 1,048,576 | ~786,000 | ~1,573 | |
| Gemini 2.5 Flash-Lite | 1,048,576 | ~786,000 | ~1,573 | |
| GPT-5.6 Sol | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-5.6 Terra | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-5.6 Luna | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-5.5 | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-5.5 Pro | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-5.4 | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-5.4 Pro | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-4.1 | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-4.1 mini | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| GPT-4.1 nano | OpenAI | 1,047,576 | ~786,000 | ~1,571 |
| Claude Fable 5 | Anthropic | 1,000,000 | ~750,000 | ~1,500 |
| Claude Opus 5 | Anthropic | 1,000,000 | ~750,000 | ~1,500 |
| Claude Sonnet 5 | Anthropic | 1,000,000 | ~750,000 | ~1,500 |
| Claude Opus 4.8 | Anthropic | 1,000,000 | ~750,000 | ~1,500 |
| Claude Sonnet 4.6 | Anthropic | 1,000,000 | ~750,000 | ~1,500 |
| DeepSeek V4 Flash | DeepSeek | 1,000,000 | ~750,000 | ~1,500 |
| DeepSeek V4 Pro | DeepSeek | 1,000,000 | ~750,000 | ~1,500 |
| Grok 4.3 | xAI | 1,000,000 | ~750,000 | ~1,500 |
| Qwen3.7 Max | Alibaba | 1,000,000 | ~750,000 | ~1,500 |
| Qwen3.5 Plus | Alibaba | 1,000,000 | ~750,000 | ~1,500 |
| Llama 4 Maverick | Meta | 1,000,000 | ~750,000 | ~1,500 |
| Grok 4.5 | xAI | 500,000 | ~375,000 | ~750 |
| GPT-5.4 mini | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5.4 nano | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5.2 | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5.2 Pro | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5.1 | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5 | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5 mini | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5 nano | OpenAI | 400,000 | ~300,000 | ~600 |
| GPT-5 Pro | OpenAI | 400,000 | ~300,000 | ~600 |
| Mistral Medium 3.5 | Mistral | 262,144 | ~197,000 | ~393 |
| Mistral Large 3 | Mistral | 262,144 | ~197,000 | ~393 |
| Kimi K2.6 | Moonshot AI | 262,144 | ~197,000 | ~393 |
| Kimi K2 Thinking | Moonshot AI | 262,144 | ~197,000 | ~393 |
| Grok Code Fast 1 | xAI | 256,000 | ~192,000 | ~384 |
| Grok Build 0.1 | xAI | 256,000 | ~192,000 | ~384 |
| Qwen3 Max | Alibaba | 252,000 | ~189,000 | ~378 |
| o3 | OpenAI | 200,000 | ~150,000 | ~300 |
| o4-mini | OpenAI | 200,000 | ~150,000 | ~300 |
| Claude Sonnet 4.5 | Anthropic | 200,000 | ~150,000 | ~300 |
| Claude Haiku 4.5 | Anthropic | 200,000 | ~150,000 | ~300 |
| GPT-4o | OpenAI | 128,000 | ~96,000 | ~192 |
| GPT-4o mini | OpenAI | 128,000 | ~96,000 | ~192 |
| Magistral Medium | Mistral | 128,000 | ~96,000 | ~192 |
Words and pages use English-prose averages (0.75 words per token, 500 words per page); code and non-English text fit less. Check whether a specific document fits with the words to tokens converter, or paste it into the token counter for a live context bar.
FAQ
Context window FAQ
What is a context window?
The context window is the maximum number of tokens a model can process in one request — the system prompt, conversation history, retrieved documents, and the response all share it. Text beyond the window is truncated or rejected, so long-document workflows are constrained by this limit.
Which AI model has the largest context window?
In this catalog, Llama 4 Scout has the largest context window at 10,000,000 tokens — roughly 7,500,000 English words. Several other models support 1 million tokens, while most flagship models offer 200,000 to 400,000.
How many pages fit in a 200K context window?
A 200,000-token window holds roughly 150,000 English words, or about 300 pages at 500 words per page. In practice you should budget less, because the response and system instructions consume the same window.
Does a bigger context window cost more to use?
The window itself is free — you pay for the tokens you actually send. But large windows invite large prompts, and some providers bill higher per-token rates above a threshold (for example, beyond 200K tokens on certain Gemini models). Filling a 1M window can cost dollars per request at flagship rates.