FREE TOOL · NO SIGN-UP

Token counter for GPT, Claude and Gemini

Paste a prompt or document to count its tokens with the o200k and cl100k tokenizers, see characters and words, and estimate what it costs to send.

RUNS IN YOUR BROWSERNOTHING UPLOADEDNO WATERMARK
Tokens
…
Words
22
Characters
132
Chars per token
—

Exact for OpenAI models. Claude, Gemini, DeepSeek and Qwen use their own tokenizers, so their counts land near this one but are not identical. Estimate the cost of this prompt →

How to use it

Step 1
Paste or type your text.
Step 2
Pick the tokenizer: o200k for GPT-4o and newer OpenAI models, cl100k for GPT-4 and GPT-3.5.
Step 3
Read the token, word and character counts, and the estimate for other model families.

Frequently asked questions

Is the count exact?
For OpenAI models, yes: the page runs the same o200k and cl100k byte-pair encodings the models use. Claude, Gemini and other families use their own tokenizers, so their numbers are estimates based on the OpenAI count.
How many tokens is a word?
In English, a token is about four characters, so 100 tokens is roughly 75 words. Code, numbers and languages such as Chinese or Japanese use more tokens per character.
Is my text sent anywhere?
Everything runs in your browser. Nothing you paste or open is uploaded.

Build with it on SandBase

Models and APIs on one key, when the job outgrows a browser tool.

More free tools

All free tools →