ModelFarebeta
YOUR COST STUDIO

Your next AI feature.

Choose a workload. See what it could cost.

Plan my costsTry a worked example

A starting point

Support chatbot

Document summary

Code review

Example usage. Choose your own provider prices.

How it works
  1. Set your workload

    Tasks, tokens, and model calls.

  2. Choose model rates

    Use sourced prices or your own provider quote.

  3. Review your forecast

    Compare scenarios and export the assumptions.

See the planning workflow
A little clarity, before you build.MF / 01
ONE USER ACTION. MORE THAN ONE CALL.
A customer needs an answer
Your primary modelContext in. An answer out.
CALL
Retry if neededAnother attempt has a cost
Fallback modelAllow for a different model

Plan every call that makes your feature work.

Workflow illustration. Interactive tabs need JavaScript.
Prices, privacy & limitations

Prices include their provider source and retrieval date. Check the current provider quote before spending.

Estimates cover supported text API costs. Cached and batch discounts, embeddings, media, hosting, taxes, and unsupported fees are excluded.

Guest calculations stay in your browser. Saving is your choice. No inference API key is needed.

A forecast does not measure answer quality. Use the local CLI to test models on your own tasks.

Price sourcesPrivacyEvaluation guide

Estimates, not provider bills.