FAQ

Frequently asked questions

How often do you update pricing data?

We review provider rate cards monthly and whenever major launches occur (e.g. new Claude or Gemini releases). Open-source GPU assumptions follow public spot pricing from AWS, Azure, and RunPod.

Do results include taxes or regional markups?

Not yet. The calculator expresses totals in USD without regional adjustments. Factor in VAT or local surcharges separately if your billing location requires it.

Can I add my own fine-tuned or self-hosted model?

Soon. Phase two of the roadmap will ship custom model entries so you can supply your own per-token or per-hour pricing. For now, duplicate an open-source entry and adjust the hourly cost.

How should I estimate success rate?

Use historical production metrics where possible. For new projects, triangulate with manual QA runs: record how many completions meet your acceptance criteria and set the success target accordingly.

Do you store my scenario inputs?

No. All calculations run in your browser session and are not persisted server-side. CSV/PDF export will allow you to save scenarios locally once released.

Is there an API?

An API is on the long-term roadmap. Our focus right now is a polished UI that helps teams validate AI investments quickly.

Still exploring?

Jump into the calculator to model your workload, or review our token estimation guide for detailed forecasting tips.