LLM API cost calculator
467 priced modelsEstimate LLM API cost by model and token volume using live AnyRouter catalog prices. Compare credit pricing with the provider list rate and see what bring-your-own-key costs.
Cheapest models for this workload
What 1M input plus 100K output tokens costs across the twelve lowest-priced models in the catalog. Prices are the headline rate on AnyRouter.
| Model | Input / 1M | Output / 1M | Workload cost |
|---|---|---|---|
| Agents A1 | $0 | $0 | $0 |
| All MiniLM | $0 | $0 | $0 |
| Amos 1 Large | $0 | $0 | $0 |
| Amos 1 Mini | $0 | $0 | $0 |
| Amos 1 Small | $0 | $0 | $0 |
| Apodex 1.1 Mini | $0 | $0 | $0 |
| Apple Foundation Model | $0 | $0 | $0 |
| Atria Dawn | $0 | $0 | $0 |
| Atria Dawn Preview | $0 | $0 | $0 |
| Beam 1.0 | $0 | $0 | $0 |
| BGE-Large | $0 | $0 | $0 |
| Codex Mini (Latest) | $0 | $0 | $0 |
Questions
Where do the prices come from?
Every figure comes from the AnyRouter model catalog, which is generated from each upstream's published rates. The calculator has no handwritten price table, so it changes when a provider changes its price.
What does bring-your-own-key (BYOK) cost?
AnyRouter adds no per-token markup on a BYOK route. The request goes out with your own provider key and your provider account is billed at its list rate; the calculator shows that rate as the provider estimate.
Do cached tokens or long-context tiers change the result?
The estimate covers input and output tokens only. Where a model publishes a cache-read rate it is shown on the model page, and models with a long-context premium list that tier separately. Neither is folded into this simple estimate.
Are free models really $0?
The anyrouter/free route and :free SKUs deduct no credits. They unlock with a $5 top-up, the Go plan or a donated key and have no daily cap. See the free-models page for the current limits.