Agent Cost Comparator

Lowest modeled TCORaw API

Monthly successful runs0

Best cost / success$0.00

Capacity statusSet workload

Workload model

Describe one repeatable task

All values editable

Workload type Runs per day Billing days / month Input tokens / run Output tokens / run Model calls / run Tool calls / run Retry rate % Cache hit rate % Concurrent runs Developer minutes / failed run Loaded developer $ / hour

Comparison lanes

Make unlike billing models comparable

Monthly TCO

Cash plus estimated labor

Formula audit

See every multiplier

Sensitivity

Low, base, and high utilization

Capacity planner

Will the lane keep up?

Requests / minute limitAverage queue secondsRun duration minutes

Decision weights

What matters besides price?

Operational checks

Turn unknowns into explicit work

Synthetic code-review example

Four model calls, eight tool steps, one successful review

This worked scenario demonstrates the calculation only. It does not reproduce or validate any provider claim from the source post.

Caveated recommendation

Compare the whole agent cost, not just tokens.