Tools🇰🇷 한국어

AI Token Counter

Count tokens for GPT, Claude and Gemini, then estimate your API cost — instantly and privately

0
GPT-4o / GPT-5
0
GPT-4 / 3.5
0
Characters
0
Words
Your Text / Prompt
Estimated Input Cost
ModelInput / 1MOutput / 1MThis prompt
Cost shown is for this text used as input tokens. Output tokens are billed separately and are usually 3–6× more expensive. Prices are reference values for June 2026 and may change — always confirm on each provider's official pricing page.
FAQ
How accurate is this AI token counter?
Counts for GPT models use the official OpenAI tokenizers (o200k_base for GPT-4o/GPT-5, cl100k_base for GPT-4 and GPT-3.5), so they match the API exactly. Claude and Gemini use their own tokenizers that are not public in the browser, so their counts are close estimates based on tiktoken.
Why does the same text use more tokens in Korean or other non-English languages?
Tokenizers were trained mostly on English, so English words often map to a single token. Korean, Japanese, Chinese, Arabic and emoji are split into many smaller tokens — sometimes 2 to 3 tokens per character — which raises both your token count and your API cost.
Is my text sent to a server?
No. The tokenizer runs entirely in your browser. Your prompt never leaves your device, so it is safe to paste private or proprietary text.
How many words is 1,000 tokens?
For English, 1,000 tokens is roughly 750 words or about 4,000 characters. The exact ratio depends on the language and content.
Advertisement (AdSense)

What is an AI token and why does it matter?

Large language models like GPT, Claude and Gemini do not read text as words or letters — they break it into tokens, small chunks that are often a few characters long. A token is roughly four characters or 0.75 words of English. Every API call is billed per token, and every model has a maximum context window measured in tokens, so knowing your token count tells you both how much a request will cost and whether it will fit.

How to use this token counter

Paste or type your prompt into the box above. The counter instantly shows the token count for the GPT-4o/GPT-5 tokenizer (o200k_base) and the older GPT-4/GPT-3.5 tokenizer (cl100k_base), along with character and word counts. The cost table then estimates what this text would cost as input across popular models. Everything runs locally in your browser using the real tiktoken vocabulary, so your text stays private.

Why non-English text costs more

Because tokenizers are trained mostly on English, languages such as Korean, Japanese and Chinese are split into many more tokens for the same meaning. A sentence that is 10 tokens in English can easily be 20–30 tokens in Korean. If you build apps for non-English users, this counter helps you see the real cost before you ship.

Share