Token Counter
Paste any text below to estimate how many tokens it uses on Claude and GPT-class models, plus what that input would cost at current API prices.
0
Estimated tokens
0
Words
0
Characters
0
Lines
Estimated input cost for this text
| Model | Input price / 1M tokens | Cost for this text |
|---|
How the estimate works: modern LLM tokenizers average roughly 3.6 characters per token for English prose, and about 1.3 tokens per word once punctuation and whitespace are counted. This tool blends both signals (characters / 3.6 and words x 1.33 plus punctuation weight) into one estimate. For typical English text it lands within about 10-15% of the true count; code, non-English text, and unusual formatting can vary more. For exact numbers, use the provider's token counting API. Claude prices shown are the published Anthropic API rates; see the LLM Price Calculator for full pricing and other providers.