Paste a prompt, support macro, or document and instantly see how much of each model's context window it consumes — plus the input cost to send it once.
100% free · no signup · browser-local
Share on XPaste text or switch to token mode if you already counted tokens.
Characters
329
Words
55
Estimated tokens
83
Pages (500w)
0.1
Estimate uses ~4 characters per token. For exact BPE counts, open the Tokenizer Diff Arena.
Cost to send once
$0.0002
Cost to fill window
$2.00
single request, input only
Pages
0.1
500 words each
Novels
0.00
90k words
Chat messages
3.3
~25 tokens each
Code lines
8.3
~10 tokens each
Sorted by context window. A red bar means the content does not fit.
| Model | Context | Usage | Remaining | Cost once | Fits? |
|---|---|---|---|---|---|
| GPT-6 AstraOpenAI | 1.1M | 1M | $0.0008 | Yes | |
| GPT-6.1 SolOpenAI | 1.1M | 1M | $0.0002 | Yes | |
| GPT-6 LunaOpenAI | 1.1M | 1M | $0.0000 | Yes | |
| Claude Fable 5.1Anthropic | 1M | 999.9K | $0.0008 | Yes | |
| Claude Opus 5.5Anthropic | 1M | 999.9K | $0.0003 | Yes | |
| Claude Sonnet 5.5Anthropic | 1M | 999.9K | $0.0002 | Yes | |
| Claude Haiku 5.5Anthropic | 1M | 999.9K | $0.0000 | Yes | |
| Gemini 4 ArgonGoogle | 1M | 999.9K | $0.0003 | Yes | |
| Gemini 3.8 FlashGoogle | 1M | 999.9K | $0.0001 | Yes | |
| Gemini 3.5 Flash-LiteGoogle | 1M | 999.9K | $0.0000 | Yes | |
| Gemini 3.1 Pro (Preview)Google | 1M | 999.9K | $0.0002 | Yes | |
| DeepSeek FlashDeepSeek | 1M | 999.9K | $0.0000 | Yes | |
| DeepSeek V4 ProDeepSeek | 1M | 999.9K | $0.0001 | Yes | |
| Grok 4.3xAI | 1M | 999.9K | $0.0001 | Yes | |
| GPT-5.5OpenAI | 922K | 921.9K | $0.0004 | Yes | |
| Grok 4.7xAI | 500K | 499.9K | $0.0002 | Yes | |
| GPT-5.4 MiniOpenAI | 400K | 399.9K | $0.0001 | Yes | |
| GPT-4o (Omni)OpenAI | 128K | 127.9K | $0.0002 | Yes | |
| GPT-4o MiniOpenAI | 128K | 127.9K | $0.0000 | Yes | |
| Llama 3.3 70B (Groq/Together)Meta | 128K | 127.9K | $0.0000 | Yes | |
| Qwen 2.5 72BAlibaba | 128K | 127.9K | $0.0000 | Yes | |
| Mistral Large 2Mistral | 128K | 127.9K | $0.0002 | Yes | |
| Command R+ (Cohere)Cohere | 128K | 127.9K | $0.0002 | Yes |
Context limits count both the prompt and the generated answer. Leave room for output, conversation history, and retrieved chunks — not just the document you paste here.
A model with a 1M-token window can ingest an entire book, but you pay for every input token on every request. If a support bot sends a 20,000-token knowledge base on each message, context size — not the per-token headline rate — becomes the dominant cost.
Use this visualizer to find the smallest window that fits your content. Then price the request in the Token Price Converter or the full cost calculator.
For long conversations, budget for history growth. A 10-turn chat re-sends earlier turns every time unless you summarize or cache them.
Keep input below 50–70% of the window so the model has room to reason and answer. If you consistently exceed 75%, switch to retrieval or add a summarization step.
Yes. The limit covers the full request: system prompt, conversation history, retrieved documents, and the generated output. If you fill the window with input, the model has no room left to answer.
Each provider uses a different tokenizer. This tool uses a fast ~4-characters-per-token estimate. For exact counts per tokenizer, use the Tokenizer Diff Arena.
Chunk and retrieve only relevant sections, summarize older conversation turns, shorten the system prompt, or move to a model with a larger context window. Bigger context is not automatically cheaper — input tokens are billed on every call.
No. Large windows cost more per request and can suffer from “lost in the middle” recall. Retrieval plus a focused prompt is often cheaper and more accurate than sending an entire corpus.
Token counts are estimates based on character density. Exact tokenizer counts vary by provider and content type.
See how OpenAI, Claude, Llama, and DeepSeek tokenizers split the same prompt.
Strip filler, boilerplate, and duplicate context to cut prompt tokens by up to 35%.
Assemble a structured prompt from 20 original templates, audit the brief, and price it at your volume.
We'll email you when model prices change or new tools launch. Unsubscribe anytime.