Verified provider pricing

Estimate your AI API bill.

Choose the kind of work and its approximate volume. You will get a useful monthly estimate before touching token ratios, caching, or batch pricing.

Loading pricing date… Published equations · no paid ranking

Current estimateLoading current prices…See the breakdown ↓

What are you trying to estimate?

Pick the closest workload. The assumptions stay visible and editable.

Chat or supportMessages per day Document search / RAGQuestions per day Agent or automationRuns per day

Messages per dayAbout 3,000 messages each month.

Model scopeChoose one model for a focused bill, or compare the market.

Refine the estimateExact tokens, cache, batch, and long contextOptional

Production assumptions

Applied to the estimate

Input served from cache0%

Uses the documented cached-input rate when a model publishes one. Cache and batch discounts do not stack.

Extra billed reasoning output0%

Adds a transparent percentage above visible output tokens.

Asynchronous batchUses a documented batch price or multiplier for the active price tier. Long-context pricingThe threshold varies by model and is shown in each result.

Loading workload assumptions…

Also useful

→ Token counter — count a real prompt before you price it → Model reference — compare price, context, and sourced benchmarks → Model recommender — guided questions → Price per use case — when cost changes everything → How to cut your token bill

Frequently asked questions

What's the cheapest AI API?

It depends on the ratio of input to output in your workload. The calculator sorts the current published prices for the usage you enter; records without a complete API price are listed separately and never treated as free.

How much does Claude cost per month?

Depends on which tier. Enter your actual input, output, caching, and reasoning-token assumptions above for a model-by-model figure.

Does this include cached pricing?

Yes. Set the repeated-input percentage and the calculator applies each model's documented cached-input rate when available. Batch and cache discounts are not stacked.

Is the pricing up to date?

Prices come from the shared benchr model index, verified against official provider pricing pages and recorded in the changelog. Provider prices can change without notice.