Updated July 1, 2026
Token estimates reference
Language models bill and measure by tokens, not words. These rules of thumb help you estimate before sending text. They are approximate: the true count depends on the model's tokenizer.
English text rules of thumb
| Unit | Approx tokens | Note |
|---|---|---|
| 1 token | ~4 characters | Common English text |
| 1 token | ~0.75 words | So 1 word is ~1.3 tokens |
| 100 tokens | ~75 words | A short paragraph |
| 1,000 tokens | ~750 words | About 1.5 pages |
| 1 page (~500 words) | ~650 tokens | Double-spaced |
Input and output both countThe context window and billing include your prompt and the model's reply. Budget tokens for the response, not just the input.
Code and other languages differSource code, punctuation-heavy text, and non-English scripts often use more tokens per character than plain English. Treat these numbers as estimates, not exact counts.
References
Questions
How many tokens is one word?
About 1.3 tokens per word of English on average, or equivalently one token per 0.75 words. Longer and rarer words split into more tokens.
How many tokens are in a page?
A typical page of around 500 words is roughly 650 tokens. This varies with formatting, language, and the specific tokenizer.