Context window checker
Will It Fit?
Paste your prompt and see instantly whether it fits inside each model's context window — before you hit the API. Everything is counted in your browser; your text never leaves this page.
Your text
14 words · 86 characters
Your text
Counting...
tokens (+2,000 reserved)
Fits in
... / 33
models checked
Smallest window that fits
Pending count
Waiting for token count
| Model | Context window | Usage | Status |
|---|---|---|---|
Claude Haiku 4.5 Anthropic | 200K | ... | Pending count |
GPT-5.4 Mini OpenAI | 400K | ... | Pending count |
GPT-5.4 Nano OpenAI | 400K | ... | Pending count |
Claude Fable 5.1 Anthropic | 1M | ... | Pending count |
Claude Opus 5.5 Anthropic | 1M | ... | Pending count |
Claude Mythos 5.1 (limited availability) Anthropic | 1M | ... | Pending count |
Claude Fable 5 Anthropic | 1M | ... | Pending count |
Claude Sonnet 5 Anthropic | 1M | ... | Pending count |
Claude Mythos 5 (limited availability) Anthropic | 1M | ... | Pending count |
Claude Opus 5 Anthropic | 1M | ... | Pending count |
Claude Opus 4.8 Anthropic | 1M | ... | Pending count |
Claude Sonnet 4.6 Anthropic | 1M | ... | Pending count |
Gemini 3.8 Flash | 1,048K | ... | Pending count |
Gemini 3.7 Flash | 1,048K | ... | Pending count |
Gemini 3.6 Flash | 1,048K | ... | Pending count |
Gemini 3.5 Flash | 1,048K | ... | Pending count |
Gemini 3.5 Flash-Lite | 1,048K | ... | Pending count |
Gemini 3.1 Pro Preview | 1,048K | ... | Pending count |
Gemini 3.1 Flash-Lite | 1,048K | ... | Pending count |
Gemini 2.5 Pro | 1,048K | ... | Pending count |
Gemini 2.5 Flash | 1,048K | ... | Pending count |
Gemini 2.5 Flash-Lite | 1,048K | ... | Pending count |
GPT-6.1 Sol OpenAI | 1,050K | ... | Pending count |
GPT-6 Astra OpenAI | 1,050K | ... | Pending count |
GPT-6 Sol OpenAI | 1,050K | ... | Pending count |
GPT-6 Luna OpenAI | 1,050K | ... | Pending count |
GPT-5.6 Sol OpenAI | 1,050K | ... | Pending count |
GPT-5.6 Terra OpenAI | 1,050K | ... | Pending count |
GPT-5.6 Luna OpenAI | 1,050K | ... | Pending count |
GPT-5.5 OpenAI | 1,050K | ... | Pending count |
GPT-5.5 Pro OpenAI | 1,050K | ... | Pending count |
GPT-5.4 OpenAI | 1,050K | ... | Pending count |
GPT-5.4 Pro OpenAI | 1,050K | ... | Pending count |
Counts use a reference tokenizer. Each provider may split text differently; message formatting, system prompts, and tool definitions add overhead. Window usage includes your reserved output tokens, but does not validate model-specific output limits.
Context window FAQ
Will my prompt fit in the context window?
Paste your text and the checker adds a reference token count to your reserved output, then compares that total against the listed context windows. Wait for counting to finish. Provider tokenizers and request overhead can change the actual total, so leave a margin.
What counts toward the context window?
Everything the model sees: your system prompt, conversation history, the text you paste, plus the tokens you reserve for the answer. Use the reserve slider to leave room for the reply.
Why do token counts differ between models?
Each model family uses its own tokenizer, so the same text splits into slightly different token counts. This checker uses a GPT-style tokenizer as a close estimate; Claude and Gemini counts may vary a little.
What happens if my text does not fit?
Most APIs reject the request with a context-length error. Some truncate the oldest content instead, which can silently remove your system prompt or early conversation turns.
How many tokens should I reserve for the output?
It depends on the task. A short classification needs a few hundred; a long summary or code generation can need several thousand. The default reserve is 2,000 tokens.
Does this tool upload my text?
No. Token counting runs in a Web Worker inside your browser. Your text never leaves your device.