Tokens to words calculator
Quick conversions for budgeting and sanity checks. These are rule-of-thumb ratios, clearly labeled as such: when the number matters, paste your actual text into a real counter.
How the conversion works
English prose averages roughly 0.75 words per token across modern tokenizers, which is where these conversions come from. The ratio is a population average, not a law:
- Code runs more tokens per word, because symbols, indentation, and identifiers each cost tokens.
- Non-English text usually costs more tokens per word on Western tokenizers, and CJK text can cost several tokens per character. Tokenizers with large multilingual vocabularies, like Kimi's, narrow this gap.
- Different providers tokenize differently, so the same text produces different counts on GPT, Claude, Gemini, Kimi, and DeepSeek. Claude counts even differ between Claude generations.
The conversions people actually look up
| Tokens | ≈ English words | Rough mental model |
|---|---|---|
| 1,000 | 750 | A long email |
| 10,000 | 7,500 | A long-form article |
| 100,000 | 75,000 | A short novel |
| 1,000,000 | 750,000 | Roughly ten novels, or a large codebase's docs |
At ~0.75 words per token, English prose. Verified against the standard industry rule of thumb, 2026-08-10.
Need the real number? The OpenAI counter runs the actual tokenizer in your browser, and the Claude counter uses Anthropic's exact counting endpoint. Background reading: what is a token?