How many tokens is your text — and what those tokens cost per month.
Every LLM meters tokens, and the word-to-token ratio quietly inflates code and non-English text. Convert token counts to a real bill on the LLM token cost calculator, and check they fit the model window on the context window calculator.
The Words to Tokens Calculator estimates how many tokens your text will use and what that will cost to process through a language-model API. You enter the number of words in your text, how many requests you send per month, and the provider's price per million tokens. The tool converts your word count into an approximate token count, then multiplies by monthly volume and the per-million rate to project your recurring bill. The main cost drivers are simple: longer text, more requests, and a higher per-token price each push the total up in direct proportion.
The key thing to watch is that tokens are not words — most models split text into sub-word pieces, so a token count usually runs somewhat higher than the word count, and the ratio shifts with punctuation, formatting, and language. Treat the output as a planning estimate, not an exact invoice, and confirm your provider's current rate. The practical lever is text length: trimming prompts and repeated context lowers tokens on every request, and because that saving is multiplied across your monthly volume, small reductions compound into a meaningfully smaller bill.
As a rule of thumb, one English word is about 1.33 tokens, and 100 tokens is roughly 75 words. Put differently, four characters is about one token. It is only an average — punctuation, numbers, code and non-English scripts split into more tokens per word.
Because APIs bill per token, not per word or per request. Your monthly spend is total tokens (input plus output, across every call) times the per-million price. A small per-request text becomes a large bill once you multiply by traffic, so estimating tokens up front is how you size the budget.
📚 New here? Learn how LLM API pricing works →
Cost at different volumes:
| Volume | Monthly | Yearly |
|---|