AI Token Counter: Count Tokens for GPT, Claude and Gemini

How to use the AI Token Counter

1. 1Paste or type your prompt, or open a text file.
2. 2Read the token count for each model family as you type.
3. 3Turn on "Show tokens" to see how the text is split into tokens.

Frequently asked questions

What is a token?

A token is a piece of text a language model reads as one unit: often a whole short word, part of a longer word, a space or a punctuation mark. In English, one token is about four characters or three quarters of a word.

Are the GPT counts exact?

Yes. GPT-5, GPT-4.1, GPT-4o and the o-series share one tokenizer, and GPT-4 and GPT-3.5 Turbo share another. Both are public, so the counts match what the model sees for your text. Chat messages add a few tokens each for formatting.

Why are Claude and Gemini counts estimates?

Anthropic and Google don't publish tokenizers that can run offline; they count tokens only through their own services. So the page estimates from a similar public tokenizer. Expect the real count to differ by roughly 10 to 20 percent, more for code or languages other than English.

What is a context window?

It's the most tokens a model can handle at once, counting your prompt and its reply. The bars show how much of each model's window your text would use.

Why does the same text give different counts for different models?

Each tokenizer has its own vocabulary. Newer ones have bigger vocabularies, so they usually need fewer tokens, especially for languages other than English and for emoji.

Is my text sent anywhere?

No. Counting happens on your device and your text never leaves it, so it's safe for private prompts and documents.