Estimate the infrastructure and service cost of one image generation task. Combine compute time, API or model charges, storage, and retry overhead to see both direct cost and a practical loaded cost per successful task.
Direct cost = GPU seconds ÷ 3,600 × GPU hourly price + service charges + storage. Loaded cost = direct cost ÷ (1 − retry rate).
The main result is the variable cost of a successful task after expected repeated attempts. It excludes fixed engineering, moderation labor, taxes, and general business overhead unless entered as a service charge.
Planning estimate only. Validate the result against production telemetry and current provider pricing before committing capacity or budget.
12 GPU-sec, $1.80/GPU-hour, $0.018 service fees, $0.004 storage, and 6% retry overhead produce a loaded cost of about $0.03 per successful image.
Use recent production telemetry or a representative benchmark on the same model, hardware, input size, and quality settings.
No. It focuses on the entered operating assumptions; add fixed staffing, development, and platform overhead separately when building a full budget.
Use the observed fraction of attempts that must be repeated, and avoid adding the same retry overhead to more than one input.
A vendor maximum is a useful ceiling, but measured sustained performance is safer for capacity planning.
Input size, batching, model version, hardware, cold starts, safety checks, and traffic bursts can all change actual results.
| Input | Purpose |
|---|---|
| Workload values | Define the size and frequency of the task. |
| Performance values | Describe measured processing behavior. |
| Planning assumptions | Add utilization, retries, price, or headroom where applicable. |