Context window checker
Does my document fit? This free context window checker tells you whether your document fits in Claude, GPT, and Gemini. Describe it by pages, slides, emails, or lines of code, with nothing to paste.
Context windows checked against provider documentation on October 21, 2026. Sizes are estimates. Runs in your browser.
What are you sending?
Describe your content by size, with no need to paste it. Add more than one kind if the request mixes them. Sizes are rules of thumb for English text.
A single-spaced page is about 500 words.
The reply shares the window with your input.
Does it fit?
Estimated size
About 201,000 tokens
Likely range: 160,800 to 241,200, plus 2,000 for the response. Fit is checked against the high end.
It fits in 12 of 13 models.
| Model | Window | Result |
|---|---|---|
| GPT-6 AstraOpenAI · Uses 23% of the window, 806,800 tokens to spare | 1.05M tokens | Fits easily |
| GPT-6.1 SolOpenAI · Uses 23% of the window, 806,800 tokens to spare | 1.05M tokens | Fits easily |
| GPT-5.5OpenAI · Uses 23% of the window, 806,800 tokens to spare | 1.05M tokens | Fits easily |
| GPT-6 LunaOpenAI · Uses 23% of the window, 806,800 tokens to spare | 1.05M tokens | Fits easily |
| Gemini 3.1 Pro (preview)Google · Uses 23% of the window, 805,376 tokens to spareA higher per-token price applies at this size. | 1.05M tokens | Fits easily |
| Gemini 3.8 FlashGoogle · Uses 23% of the window, 805,376 tokens to spare | 1.05M tokens | Fits easily |
| Gemini 3.5 Flash-LiteGoogle · Uses 23% of the window, 805,376 tokens to spare | 1.05M tokens | Fits easily |
| Claude Fable 5.1Anthropic · Uses 24% of the window, 756,800 tokens to spare | 1M tokens | Fits easily |
| Claude Opus 5.5Anthropic · Uses 24% of the window, 756,800 tokens to spare | 1M tokens | Fits easily |
| Claude Sonnet 5.5Anthropic · Uses 24% of the window, 756,800 tokens to spare | 1M tokens | Fits easily |
| Claude Haiku 5.5Anthropic · Uses 24% of the window, 756,800 tokens to spareA higher per-token price applies at this size. | 1M tokens | Fits easily |
| Gemini 4 ArgonGoogle · Uses 24% of the window, 756,800 tokens to spare | 1M tokens | Fits easily |
| Claude Haiku 4.5Anthropic · Needs about 2 requests | 200K tokens | Too large |
These are estimates. Real token counts vary by model and by the text itself, and images, tables and scanned pages can take more than plain text. Fitting in the window doesn’t guarantee the model uses every part of a long input equally well, so test on your own documents.
What fits in each context window
Approximate capacity of each window for English text, before leaving room for a response.
| Window | About | Models |
|---|---|---|
| 200K tokens | 150,000 words, or 299 pages | Claude Haiku 4.5 |
| 1M tokens | 750,000 words, or 1,493 pages | Claude Fable 5.1, Claude Opus 5.5, Claude Sonnet 5.5, Claude Haiku 5.5, Gemini 4 Argon |
| 1.05M tokens | 787,500 words, or 1,567 pages | GPT-6 Astra, GPT-6.1 Sol, GPT-5.5, GPT-6 Luna, Gemini 3.1 Pro (preview), Gemini 3.8 Flash, Gemini 3.5 Flash-Lite |
How to use it
- Describe your content. Pick what you have, such as pages, slides, or hours of transcript, and how much. Add more than one kind if the request mixes them.
- Reserve room for the answer. Say how long a response you expect. It shares the window with your input.
- Read the result. See which models fit easily, which are tight, and how many requests the rest would need.
Context windows, explained
- What is a context window?
- The most text a model can handle in a single request. Your instructions, your documents, the conversation so far, and the reply the model writes all count toward it. If the total is too large, the request fails or the oldest content is dropped.
- Does my document fit in the context window?
- Describe it above by pages, slides, emails, transcript hours, or lines of code and the checker estimates its size in tokens and compares it with every model’s window. A model is shown as fitting only if the high end of the estimate, plus room for the response, fits.
- How many pages fit in a context window?
- The table on this page converts each window into pages and words. As a rule of thumb, one token is about three quarters of an English word, and a single-spaced page is about 500 words, so a page is roughly 670 tokens.
- Why does the response need room too?
- Input and output share the same window. A model that must write a long summary or a long piece of code needs space left for it, so reserve room for the response you expect before you decide whether a document fits.
- What if my document doesn’t fit?
- Split it into chunks and process them separately, summarize in stages, or retrieve only the relevant passages instead of sending everything. The checker shows how many requests a split would take for each model.
- Is a bigger window always better?
- No. Longer inputs cost more, some models charge a higher rate above a certain prompt length, and very long inputs can make it harder for a model to use every detail well. Send only what the task needs.
- How accurate is the estimate?
- It’s a rule of thumb for English text, shown as a range because each provider counts tokens differently. Images, scanned pages, tables, code, and other languages usually take more. Context windows were checked against provider documentation on October 21, 2026. For exact numbers, use the provider’s own token counter.
I’m Courtney Stephenson, an AI strategy and implementation leader, and I built this tool. Get in touch if you’re planning around context limits. To count tokens in text you already have, try the token counter.
Want a second opinion on your results?
Book a free 15-minute call and we’ll talk through what this means for your team.