Powerful cloud native instances featuring full passthrough of one or more NVIDIA B200 GPUs.
The NVIDIA B200 is a next-generation data center GPU based on the NVIDIA Blackwell architecture, engineered for large-scale AI training, inference, and high-performance computing workloads. It features advanced Tensor Cores, high-bandwidth memory (HBM3e), and enhanced interconnect technologies that deliver significant performance and efficiency improvements over previous generations. Optimized for generative AI and trillion-parameter models, the B200 provides exceptional scalability, memory capacity, and throughput. Its cutting-edge architecture enables organizations to accelerate complex AI applications while reducing infrastructure costs and energy consumption.
| Name | NVMe Storage | Memory | vCPUs | Transfer | Price | Regions |
|---|---|---|---|---|---|---|
G4-gpu-b200-240m-24c
| 4800 GiB | 240 GB | 24 vCPUs | 60 TB | $8.3562 /hr $6 100 /mo | 15 |
G4-gpu-b200-480m-48c
| 9600 GiB | 480 GB | 48 vCPUs | 70 TB | $11.6438 /hr $8 500 /mo | 15 |
G4-gpu-b200-480m-48c
| 9600 GiB | 480 GB | 48 vCPUs | 70 TB | $17.1233 /hr $12 500 /mo | 15 |
G4-gpu-b200-960m-96c
| 19200 GiB | 960 GB | 96 vCPUs | 80 TB | $23.2877 /hr $17 000 /mo | 15 |
G4-gpu-b200-960m-96c
| 19200 GiB | 960 GB | 96 vCPUs | 80 TB | $33.5616 /hr $24 500 /mo | 15 |
G4-gpu-b200-1920m-192c
| 38400 GiB | 1920 GB | 192 vCPUs | 90 TB | $66.4384 /hr $48 500 /mo | 15 |