LLMTokenBase

Token counter for tiktoken, OpenAI, Claude, and more.

API bills and context windows speak tokens, not pages. This is the live token counter for OpenAI-style tiktoken encodings, plus approximate Claude, ChatGPT, Gemini, GPT, and Anthropic length checks — switch modes in one paste box. Approximate modes stay labeled; your text never leaves the browser.

Exact mode: browser encoding o200k_base. Common model IDs that use this encoding: gpt-4o, gpt-4o-mini, gpt-4.1, gpt-4.1-mini, gpt-5, gpt-5-mini, o3, o3-mini, o4-mini. Newer IDs without a mapping do not rename this mode.

Tokens0
Words0
Chars0
≈ $0.0000 cost (estimate)

Good uses for this page

01

Check a prompt or completion before it hits a context limit.

02

Compare draft length while you edit, without uploading the text.

03

Get a planning estimate before you spend API budget.

04

Use OpenAI-style modes when you need tiktoken-class length; use Claude or Gemini modes when you only need a labeled estimate.

How this is counted

For OpenAI-style modes we use tiktoken-compatible encodings in the browser when available. For Claude- and Gemini-oriented modes we use documented approximation rules and mark the result as an estimate. The badge is intentional.

Characters are the raw string length. Words are a whitespace split only (spaces, tabs, newlines) — not linguistic word segmentation. Continuous Chinese with few spaces can show a low word count while characters stay high; that is expected. Chat formatting, system messages, and tool wrappers can still add tokens beyond this page, so treat the number as planning guidance, not a receipt. Read the full methodology →

Need a closer vendor match? Switch modes on this page. Exact billing always comes from the provider API or console.

Exact modes in the browser use only these encodings: o200k_base, cl100k_base. Model IDs that currently map to those encodings: gpt-4o, gpt-4o-mini, gpt-4.1, gpt-4.1-mini, gpt-4, gpt-4-turbo, gpt-3.5-turbo, gpt-5, gpt-5-mini, o3, o3-mini, o4-mini. Watched IDs still without a browser encoding (listed here only; exact mode names stay encoding-based, e.g. OpenAI (o200k), not marketing names like GPT-5.6): gpt-5.6, gpt-5.6-sol, gpt-5.6-terra, gpt-5.6-luna, gpt-5.4, gpt-5.4-mini.

Related tools

Questions

No. Counting runs in your browser. If you close the tab, what you pasted goes with it.

Estimates only. Not an official tokenizer from OpenAI, Anthropic, or Google. Verify with your provider before billing decisions.