LLM Token Counter
Count tokens for GPT-5, GPT-4.1, GPT-4o and other OpenAI models with the same tokenizer they use, and see exactly how your prompt is split. Runs 100% in your browser; your text is never sent to OpenAI or to us.
Models: GPT-5, GPT-5 mini/nano, GPT-4.1, GPT-4o, GPT-4o mini, o1, o3, o4-mini
Tokens
73
estimating…
Characters
291
characters
Words
51
words
Chars per token
–
average
Prompt or text
How the text is tokenized
Loading tokenizer…
About tokens
- LLMs read and bill text in tokens, chunks that are often a whole short word, part of a longer word, a space plus a word, or punctuation.
- For English, one token is roughly 4 characters or ¾ of a word. Code, numbers, non-English text and emoji usually need more tokens per character.
- o200k_base (GPT-4o and newer) has twice the vocabulary of cl100k_base (GPT-4, GPT-3.5), so the same text often needs fewer tokens, especially in non-English languages.
- Chat APIs add a few tokens of formatting per message, so a request uses slightly more tokens than its text alone.
- Claude, Gemini and Llama use their own tokenizers; their counts for the same text differ, so use these numbers only as a rough guide for those models. To count words and characters instead, use the word counter.