Cheapest on-demand rate per card, 6 cards.
| GPU | On-demand / GPU-hour |
|---|---|
| NVIDIA L40S Everywhere Inference | $1.22 read |
| NVIDIA A100 80GB Everywhere Inference | $1.42 read |
| NVIDIA H100 80GB (form factor not stated) Everywhere Inference | $1.72 read |
| NVIDIA H100 SXM Bare Metal GPU Clusters | $3.07 read |
| NVIDIA H200 Bare Metal GPU Clusters | $3.58 read |
| NVIDIA B300 Bare Metal GPU Clusters | $6.76 read |