A token is a piece of text that an LLM processes. It can be a word, part of a word, or a character. 'Hello world' is typically 2 tokens. 1 token ≈ 4 characters in English.

Is this token counter free?

Yes, 100% free. Count tokens for unlimited text.

Is counting accurate?

For GPT models, it uses the cl100k_base tokenizer (same as OpenAI). For other models, it provides estimates based on similar tokenizer ratios.

✓ tiktoken accurate✓ Multi-model✓ Free forever

Free Token Counter for LLMs

Count tokens for GPT-4, Claude, Gemini, and Llama. Uses the same tokenizer as OpenAI.

How to use Token Counter

Three steps. Instant results.

Step 1

Paste your text or prompt in the input area.

Step 2

See real-time token counts for GPT-4, Claude, Gemini, and Llama.

Step 3

Use the counts to manage context windows and estimate API costs.

How Token Counter works

This tool uses js-tiktoken (the JavaScript port of OpenAI's tiktoken) with the cl100k_base encoding — the same tokenizer used by GPT-4 and GPT-4o. It splits your text into BPE (Byte Pair Encoding) tokens, counting the exact number of tokens that the model will process. For other models like Claude and Gemini, it applies ratio adjustments based on known tokenizer differences. The tokenizer loads once and runs entirely in your browser.