Calculator

Scenario Studio

Describe your real usage — seats, requests, prompt size, caching, retries, a fallback model, gateway fees, tax — and get the monthly API bill, broken down line by line and compared to a flat subscription. Save scenarios locally and share them by link.

Seats / users10

Requests / user / day200

Work days / month22

Per request

Input tokens12,000

Output tokens3,000

Models

Primary model

Add a fallback / cheaper-route model

Prompt caching

Model a cached prompt prefix

Adjustments

Retry overhead1.10×

Extra tokens from retries / re-runs

Batch discount0%

Async/batch API price cut, if applicable

Gateway fee

Gateway flat fee$0/mo

Fixed monthly proxy / gateway subscription, on top of the % markup

Tax / VAT

Compare to a subscription

Estimated monthly API cost

$3,920

/ month

$392/seat · $0.089/request · 44,000 req/mo

Cost component	Monthly
Base API cost	$3,564
Retries (+10%)	+$356
Total	$3,920

Order of operations: cache-aware base → retries → batch discount → gateway fee → tax. 30-day-equivalent month from your days/seat inputs. Estimates from published per-token rates; real spend varies with prefix stability and traffic mix.

How the estimate is built

The base cost blends your primary and (optional) fallback model by traffic share, applying prompt-cache rates when enabled. Then adjustments stack in order: retries inflate the token bill, a batch discount cuts it, a gateway fee is added on the subtotal, and tax is applied last. The result is a per-month, per-seat, and per-request figure.

The subscription comparison reuses the same usage-runway model as the plans tool: a plan's published caps are directional (premium-request pools, rate windows), so treat the API-vs-plan gap as a budgeting guide, not a contract. Scenarios are stored only in your browser; a shared link encodes the whole scenario in the URL.