Token Counter

How to count tokens

1. Paste your prompt, document or chat history.
2. The exact token count for OpenAI's tokenizers appears instantly, along with words and characters. Choose o200k for GPT-4o, GPT-4.1, the o-series and GPT-5, or cl100k for GPT-4 and GPT-3.5.
3. Tick Show tokens to see how the text is split, with each token highlighted. Estimates for Claude, Gemini, Llama and other model families are shown below.

What tokens are and why they matter

Large language models do not read letters or words; they read tokens, chunks of text from a fixed vocabulary of around 100,000 to 200,000 pieces. A common English word is often one token, including its leading space, while rare words, names, numbers, emoji and code are split into several. As a rule of thumb, one token is about four characters or three-quarters of an English word, but other languages often need more tokens for the same meaning.

Token counts matter for three reasons. API pricing is per token, for both input and output. Every model has a context window, the maximum number of tokens it can consider at once, including your prompt, any documents, the chat history and the reply. And longer prompts are slower. Counting before you send helps you trim prompts, split documents into chunks that fit, and budget costs with our LLM Cost Calculator.

The OpenAI counts use the open-source gpt-tokenizer library, a faithful port of OpenAI's tiktoken, loaded once into your browser. Anthropic, Google and others do not publish browser tokenizers, so their figures are labelled estimates. Your text never leaves the page.