What is the Tokens to Words & Cost Calculator?
Large language models (LLMs) like GPT, Claude and Gemini process text in tokens rather than words. A token is a chunk of text — often a word fragment of about four characters. This calculator estimates how many words a given number of tokens represents and what those tokens cost based on a model's price per 1M tokens.
How to use it
Enter the number of tokens, the words-per-token ratio (0.75 is a good default for English prose), and, optionally, the price your provider charges per 1M tokens (leave it blank to count words only). The tool instantly returns an estimated word count and, if you entered a price, the dollar cost.
The formula explained
The word estimate uses \(\text{words} \approx \text{tokens} \times 0.75\), reflecting OpenAI's rule of thumb for English that 1 token ≈ 4 characters ≈ 0.75 words. Cost is $$\text{Cost} = \frac{\text{tokens}}{1{,}000{,}000} \times \text{price per 1M}$$ since most providers quote prices per million tokens. Adjust the ratio for other languages or code, which tend to use more tokens per word.
Worked example
Suppose you have 1,000 tokens at a ratio of 0.75 and a price of 2 USD per 1M tokens. $$\text{Words} = 1{,}000 \times 0.75 = 750 \text{ words}$$ $$\text{Cost} = \frac{1{,}000}{1{,}000{,}000} \times 2 = 0.002 \text{ USD}$$
FAQ
How accurate is the word estimate? It's an approximation. Actual tokenization varies by model and content; 0.75 words/token is a solid average for English.
Do input and output tokens cost the same? No — many providers charge different rates. Run the calculator twice with each rate to add them up.
Why do code or non-English texts differ? They often pack more tokens per word, so lower the ratio (e.g. 0.5) for a better estimate.