Token counter
Paste any text and get the exact token count from a real tokenizer — plus characters, the chars-per-token ratio, and every token shown as a colored chip with its ID. Everything runs in your browser; your text never leaves your device.
GPT/OpenAI counts are exact (o200k_base tokenizer). Anthropic, Google, and DeepSeek ship no client-side tokenizer, so their counts are estimates using the GPT tokenizer as a proxy — real counts differ. Prices/metadata last checked 2026-06-06.
What is a token?
LLMs don't read characters or words — they read tokens, sub-word chunks produced by a byte-pair-encoding (BPE) tokenizer. Common words are usually one token; rarer words, code, punctuation, and other languages split into several. You're billed per token (in and out), and the context window is measured in tokens, so token count — not character count — is what actually matters.
Why the model picker matters
Different model families use different tokenizers, so the same text can be a different number
of tokens on each. This tool uses the real o200k_base encoding (GPT-4o / GPT-5
family) for exact OpenAI counts and as a labeled proxy for other providers, whose tokenizers
aren't available in the browser.