How much will your AI project actually cost once in production? This calculator estimates the monthly bill for an artificial intelligence API based on your usage volume and compares several models side by side.
How to use it
- Enter the number of requests per month, the input and output tokens per request, and any reasoning tokens.
- Specify the portion of input served by cache if you use one, then choose euros or dollars.
- Add the models to compare, or check the option to compare all paid models. Click a column header to sort.
The principles behind the calculation
- The cost of a request is the input price multiplied by the input tokens, plus the output price multiplied by the output and reasoning tokens, all divided by one million (prices are shown per million tokens).
- The portion of input served by cache is billed at the cache read rate when the model offers one; the remainder at the normal input rate.
- Some models apply a different rate beyond a context threshold: the calculator switches automatically and flags requests that exceed the model’s context window.
- Prices are the public rates recorded daily; conversion to euros uses an exchange rate updated with them. The displayed index is the one from Artificial Analysis when it exists.
Limitations to be aware of
Cache write costs, search tools, and variable rates depending on the time are not included. The result is an order of magnitude for budgeting, not an invoice.
