Documentation
Provision single-tenant GPU nodes and clusters, verify them before use, and operate them from the API, CLI or console.
Start here
All pages
Compute
- D-02NodesWhat a node is, what you control on it, and its lifecycle from assignment to teardown.
- D-03ClustersReserved groups of nodes on a dedicated, rail-optimised fabric for distributed training.
- D-04NetworkingFabric topology, private networks, public addresses and private interconnect.
- D-05StorageLocal NVMe, block volumes and object storage, and how each is encrypted and erased.
Security
Plan identifiers
Sample dataPass these to --plan or the plan field. Availability and regions here are illustrative until the inventory service is connected.
| Plan | Accelerator | GPUs | Memory / GPU | Regions |
|---|---|---|---|---|
| cl.b300.x8 | NVIDIA B300 | 8 | 288 GB | US East |
| cl.b200.x8 | NVIDIA B200 | 8 | 180 GB | EU North, US East |
| cl.h200.x8 | NVIDIA H200 | 8 | 141 GB | EU North, EU Central, US East |
| cl.h100.x8 | NVIDIA H100 | 8 | 80 GB | EU North, EU Central, US East |
| cl.mi300x.x8 | AMD MI300X | 8 | 192 GB | US East |
| cl.rtx6000.x8 | NVIDIA RTX PRO 6000 | 8 | 96 GB | EU Central, US East |
| cl.rtx6000.x2 | NVIDIA RTX PRO 6000 | 2 | 96 GB | EU North, EU Central, US East |
| cl.l40s.x4 | NVIDIA L40S | 4 | 48 GB | EU North, EU Central, US East |
| cl.l40s.x1 | NVIDIA L40S | 1 | 48 GB | EU North, EU Central |