Pricing
Honest pricing. Per second. No surprises.
Spin up an L4 for a notebook, reserve a rack of H100s, or run everything inside your own cloud. Same control plane, same SDK.
On-demandReserved (1yr)
Developer
Pay-as-you-go for individuals and small projects.
$0.39/ GPU-hour, L4
- Credit-card billing, per-second
- L4, A10, A100 pools
- Community support on Discourse
- Free egress to S3 and your warehouse
Most popular
Team
For teams shipping models to production.
$2.49/ H100-hour
- Shared workspaces and RBAC
- Reserved pools with burst to spot
- Private VPC peering
- Slack-connect support, 4h SLA
Scale / Enterprise
Custom capacity, security, and SLAs.
Customannual contract
- Dedicated regions and isolated clusters
- BYOC, BYOK, SSO, audit logs
- 99.99% uptime SLA, named CSM
- Procurement, MSA, security reviews
Prices in USD. Spot capacity available at up to 70% off on-demand.