Start free. Pay as eval volume grows.
No minimum seats. No unexpected overages. Cancel anytime.
For individual developers testing AI evaluation.
Start free- 10,000 evaluations/month
- 1 project
- Basic scoring rubrics
- 7-day trace retention
- Community support
For teams launching AI features to production.
Get Pro- 200,000 evaluations/month
- Unlimited projects
- Custom scoring rubrics
- Hallucination detection
- 30-day trace retention
- Drift alerting
- Email support
For organizations running many AI products at scale.
Get Team- 2,000,000 evaluations/month
- Unlimited projects
- All Pro features
- Human-in-the-loop escalation
- 90-day trace retention
- SSO / SAML
- Priority support
- Custom onboarding
Feature comparison
| Feature | Dev | Pro | Team |
|---|---|---|---|
| Monthly evaluations | 10,000 | 200,000 | 2,000,000 |
| Projects | 1 | Unlimited | Unlimited |
| Custom rubrics | |||
| Hallucination detection | |||
| Drift alerting | |||
| Human escalation | |||
| Trace retention | 7 days | 30 days | 90 days |
| SSO / SAML | |||
| Support | Community | Priority |
Frequently asked questions
One evaluation is a single LLM response passed through your scoring rubrics. Failed evaluations that need a retry count separately.
Yes. Upgrades begin immediately, with the prorated difference billed. Downgrades begin when your next billing cycle starts.
Dev includes 10,000 free evaluations monthly for testing. Contact us for a full Pro or Team trial, and we can set up a 14-day extended trial.
We notify you at 80% of your limit. After that, evaluations pause until the next billing cycle or an upgrade. No surprise charges.
Yes. Our startup program supports early-stage companies, while our research program supports academic teams. Contact [email protected] with your project details.