Free tool

    Token counter

    Paste a prompt to count its tokens with OpenAI’s own tokenizers, see exactly how the text splits, and estimate Claude and Gemini. Runs entirely in your browser — no AI calls, nothing uploaded.

    [ Free · No sign-up · Runs in your browser ]
    Tool

    Count tokens

    172 characters · 27 words

    How the text splits

    You are a helpful assistant. Summarize the customer email below in three bullet points, then draft a short, friendly reply that answers their question about shipping times.

    GPT-5 / GPT-4oexact

    33tokens

    GPT-5, GPT-4.1, GPT-4o, o1, o3, o4-mini

    GPT-4 / GPT-3.5exact

    33tokens

    GPT-4, GPT-4 Turbo, GPT-3.5 Turbo, text-embedding-3

    Claude, Gemini and othersestimate

    ~43tokens

    Claude, Gemini, Llama, Mistral, DeepSeek

    What would 33 input tokens cost?
    Rules of thumb

    Tokens at a glance

    1 token ≈ 4 characters

    About 0.75 of a word in English, so 100 tokens is roughly 75 words and 1,000 tokens about a page and a half.

    Everything counts

    System instructions, earlier messages in the chat, tool definitions and the model’s reply are all tokens you pay for.

    Context windows are in tokens

    A 128k-token model holds roughly 96,000 words of input and output combined.

    FAQ

    Common questions

    What is a token?

    Language models read text as tokens: common words are one token, longer or rarer words split into several, and spaces and punctuation count too. In English a token averages about 4 characters or 0.75 of a word.

    Are these counts exact?

    For OpenAI models, yes: the tool runs OpenAI's own tokenizers (o200k_base for GPT-4o, GPT-4.1, GPT-5 and the o-series; cl100k_base for GPT-4 and GPT-3.5) in your browser. Anthropic and Google don't publish their tokenizers, so Claude and Gemini counts are estimates from text length — fine for budgeting, not for exact limits.

    Is my text sent anywhere?

    No. The tokenizer is downloaded once and runs on your device. Nothing you paste leaves the page, and no AI model is called.

    Why do I need to count tokens?

    APIs bill per token and every model has a context window measured in tokens. Counting tells you what a prompt will cost and whether it fits — the system prompt, chat history and the reply all count.

    Why do other languages use more tokens?

    Tokenizers are trained mostly on English, so other languages, code and emoji often take more tokens for the same meaning. Newer tokenizers like o200k_base are more efficient for non-English text than older ones.

    How do I turn tokens into cost?

    Multiply input tokens by the model's input price and output tokens by its output price (both per million tokens). The LLM API cost calculator does it for 34 models — follow the link under the counts to carry this number over.

    Automate it

    Every model behind one account

    Chat with GPT, Claude, Gemini and more, build agents and run workflows on MultiverseAI — or use your own API keys at no credit cost.

    100 free credits to start
    No credit card required
    Cancel anytime