LLM Token Counter
Paste a prompt, document or chat log to count tokens and compare what it costs across models.
Standard list prices per 1 million tokens, checked 2026-09-25. Edit any price to match your plan. Claude and Gemini token counts are estimates.
LLM API prices per million tokens
| Model | Input | Output | 1,000-word prompt* |
|---|---|---|---|
| Claude Fable 5.1 · Anthropic | $10.00 | $50.00 | $0.0133 |
| Claude Opus 5 · Anthropic | $5.00 | $25.00 | $0.0067 |
| Claude Sonnet 5 · Anthropic | $2.00 | $10.00 | $0.0027 |
| Claude Haiku 4.5 · Anthropic | $1.00 | $5.00 | $0.0013 |
| GPT-6 Astra · OpenAI | $10.00 | $50.00 | $0.0133 |
| GPT-6 Sol · OpenAI | $2.00 | $10.00 | $0.0027 |
| GPT-6 Luna · OpenAI | $0.10 | $0.50 | $0.0001 |
| Gemini 3.1 Pro · Google | $2.00 | $12.00 | $0.0027 |
| Gemini 3.8 Flash · Google | $0.75 | $3.75 | $0.0010 |
| Gemini 3.5 Flash-Lite · Google | $0.30 | $2.50 | $0.0004 |
*About 1,333 tokens of English input. Prices as listed by each provider on 2026-09-25; batch, cached and long-context rates differ.
How it works
Your text is split with the o200k byte-pair tokenizer that OpenAI publishes. Cost = input tokens × input price + output tokens × output price, times the number of requests. The tokenizer (about 1 MB) loads on first use and everything runs in your browser.
What is a token?
A token is the unit a language model reads and writes — usually a word piece of three to four characters in English. “Tokenization” splits text into these pieces; code, numbers and non-English text often need more tokens per word.
Are these token counts exact?
The OpenAI count uses the open-source o200k tokenizer and is exact for models built on it. Anthropic and Google use their own tokenizers, so the Claude and Gemini figures here are estimates and can differ noticeably, especially for code and non-English text. For billing-exact numbers use each provider’s token-counting endpoint.
Is my text sent anywhere?
No. Tokenization runs entirely in your browser. Nothing you paste is uploaded, logged or stored.
Why do output tokens cost more?
Generating text is more compute-intensive than reading it, so providers charge several times more per output token. Long answers — and reasoning or thinking tokens, which are billed as output — often dominate the cost of a request.