How many tokens is a word?

In English, about 0.75 words per token, or roughly 4 characters per token. So 1,000 tokens is around 750 words. Other languages and scripts usually take more tokens for the same meaning.

Why does Claude show more tokens?

Anthropic says its newer Claude models (4.7 and later) produce about 30% more tokens for the same text than its older ones, so the counter adds that. A higher count does not mean a higher bill on its own; the price per token matters just as much.

No. It is an estimate, close enough for planning prompt sizes and costs. For exact numbers, use each vendor's token-counting tool on your real prompt.

Is my text sent anywhere?

No. Counting happens in your browser and nothing you paste leaves this page.

Free tool

How many tokens is your prompt, and what does it cost?

Paste any text to count characters and words, estimate its tokens, and compare what it costs to send on Claude, GPT and Gemini. Nothing you paste leaves your browser.

Prices checked

12 October 2026, from each vendor's published pricing. Token counts are estimates.

Also try the AI cost calculator and the chatbot ROI calculator.

Your text

Characters

Words

Tokens (est.)

Claude tokens (est.)

Estimate: about 4 characters per token for English, roughly 1 token per character for non-Latin scripts and emoji, and about 30% more for newer Claude models. Exact counts differ by vendor; for billing-grade numbers, use each vendor's own token counter.

Cost to send this text as input

× requests

Claude Fable 5.1 · Flagship$0
$0 per request
Claude Opus 5.5 · Flagship$0
$0 per request
Claude Sonnet 5.5 · Mid-range$0
$0 per request
Claude Haiku 5.5 · Small & fast$0
$0 per request
GPT-6 Astra · Flagship$0
$0 per request
GPT-6.1 Sol · Mid-range$0
$0 per request
GPT-6 Luna · Small & fast$0
$0 per request
Gemini 3.1 Pro (preview) · Flagship$0
$0 per request
Gemini 3.8 Flash · Mid-range$0
$0 per request
Gemini 3.5 Flash-Lite · Small & fast$0
$0 per request

Input only, standard prices, no caching. Answers are billed on top, at five to eight times the input rate. For the full monthly picture with answers, caching and batch, use the AI cost calculator.

Tokens, briefly

Prompt size is the part of the bill you control.

Every request pays for the whole prompt: instructions, history and any documents you paste in. Trimming what you send, or caching the part that repeats, usually saves more than switching models.

How many tokens is a word?: In English, about 0.75 words per token, or roughly 4 characters per token. So 1,000 tokens is around 750 words. Other languages and scripts usually take more tokens for the same meaning.
Why does Claude show more tokens?: Anthropic says its newer Claude models (4.7 and later) produce about 30% more tokens for the same text than its older ones, so the counter adds that. A higher count does not mean a higher bill on its own; the price per token matters just as much.
Is this count exact?: No. It is an estimate, close enough for planning prompt sizes and costs. For exact numbers, use each vendor's token-counting tool on your real prompt.
Is my text sent anywhere?: No. Counting happens in your browser and nothing you paste leaves this page.
Paying more than you expected?: The AI Cost Review looks at your real usage and shows where the bill can come down without hurting the answers.

Building something with AI?

Tell us what it should do. We'll tell you which model fits, what it will really cost, and build it.

Start a project Fixed-price checks