CostUse

AI models

AI model cost comparator

Select the models and enter your workload once. Every result then uses identical volume assumptions, making the cost difference easier to inspect.

View the full rate table

Choose up to five models

Select two to five models. The resulting comparison has its own link so you can bookmark or share it.

Shared workload

Uses the input/output mix above and a 30-day month: requests per day × tokens per request × 30.

Results using the same assumptions

Choose two to five models, then select “Compare selected models.”

Shared token-volume table

A 70/30 input/output estimate at fixed token bands for every selected model.

These are public list-price estimates, not invoices. Batch, priority, long-context, cache, regional, and negotiated rates can change the result.