Turn token usage into a monthly API estimate.
Compare text-generation model routes using the latest valid pricing fields already collected by AI on Radar. Every row links back to the catalog source and shows when it was observed.
- Priced routes
- 31
- Providers
- 14
- Price source
- OpenRouter Models API
- Persistence
- No workload saved
Describe one month of usage
Use billed token counts from a representative request when possible.
Compare model routes
Lowest estimated total first. Filter after calculating.
What the estimate includes—and what it does not.
For each model, the calculator multiplies monthly uncached input tokens by the listed prompt rate, cached input tokens by the listed cache-read rate when available, output tokens by the completion rate, and requests by any listed per-request charge. The catalog reports prices per token; the page converts them to USD per one million tokens for display.
Tokenizers differ, so two models can bill different token counts for the same text. Image, audio, web-search, fine-tuning, storage, taxes, credit-purchase fees, enterprise discounts, and unlisted reasoning charges are not inferred. Treat the result as a planning estimate and verify the linked source before spending.