About Token Counter
A token is the unit of text a large language model reads and bills by, usually a word fragment of a few characters. This tool does not run a real tokenizer. It divides the number of characters by four and rounds up, a common rule of thumb for English text. Actual counts vary by model and language, and code, numbers, and non-Latin scripts typically use more tokens than this estimate suggests.
Also known as: AI Token Counter, LLM Token Counter.
How to use Token Counter
- Type or paste into the "Input text" box, or press "Use example" to try it with sample data.
- The result appears below and updates as you type.
- Copy the result with the copy button, or download it as a file where that is offered.
Questions about Token Counter
How many tokens is 1,000 words?
For ordinary English prose, a common rule of thumb is about 1,300 to 1,400 tokens per 1,000 words, or roughly 0.75 words per token. The exact figure depends on the model and the vocabulary of the text.
Why does the estimate differ from the count my API reports?
This tool uses a fixed ratio of four characters per token. Real tokenizers split text with a learned vocabulary that differs per model, and APIs also count system prompts, message formatting, and tool definitions.
Runs in your browser. Page last updated .