Pay for reasoning that ships, not seats that sit idle
Every plan includes full traceability and governance. Scale up when your agents start doing real work.
Starter
For teams piloting their first agent.
- 1 active agent
- 10K reasoning steps / mo
- Community support
- 7-day trace retention
- Standard model access
Growth
For teams running agents in production.
- Unlimited agents
- 2M reasoning steps / mo
- Priority support + SLA
- 90-day trace retention
- Full governance suite
- All frontier models
Enterprise
For regulated, mission-critical deployments.
- Dedicated capacity
- Private model hosting
- Unlimited trace retention
- Custom compliance review
- 24/7 dedicated support
- Single-tenant deployment
Compare plans
| Feature | Starter | Growth | Enterprise |
|---|---|---|---|
| Active agents | 1 | Unlimited | Unlimited |
| Reasoning steps | 10K/mo | 2M/mo | Custom |
| Reasoning trace retention | 7 days | 90 days | Unlimited |
| Policy guardrails | — | ✓ | ✓ |
| Private model hosting | — | — | ✓ |
| SSO / SCIM | — | ✓ | ✓ |
| Dedicated support | — | — | ✓ |
Pricing questions
What counts as a reasoning step?
Each discrete inference an agent makes while planning, calling a tool, or producing an output counts as one reasoning step, regardless of which underlying model handled it.
Can I mix models across plans?
Yes. Starter includes standard model access; Growth and Enterprise unlock the full frontier model catalog with automatic routing and failover.
Do you offer non-profit or academic pricing?
We offer discounted Growth-tier pricing for qualifying research and non-profit organizations. Reach out to sales to check eligibility.
What happens if I exceed my monthly reasoning steps?
We'll notify you as you approach your limit. Overages are billed at your plan's per-step rate rather than cutting off agents mid-task.