NVIDIA B200 price per hour
The cheapest on-demand NVIDIA B200 (180 GB) we found is $3.69/GPU-hr at DeepInfra (managed endpoint). Across 9 listings, prices run up to $13.00. Blackwell generation with FP4 support, aimed at high-throughput inference and training.
| Provider | $/GPU-hr | Type | Variant | Notes | |
|---|---|---|---|---|---|
| DeepInfra | $3.69 | Managed endpoint | 180GB | Dedicated deployment, billed per minute | source |
| Modal | $6.25 | Serverless | Billed per second ($0.001736/s) | source | |
| Lambda | $6.69 | Cloud VM | SXM6 | Instance size not stated on page | source |
| RunPod | $7.99 | Cloud VM | SXM | source | |
| Together AI | $8.19 | Cloud VM | HGX cluster | Preemptible $4.09; 91–180 day reserved $6.79 | source |
| Nebius | $8.50 | Cloud VM | HGX | Up to 35% less with multi-month reservations | source |
| CoreWeave | $8.60 | Cloud VM | HGX 8x | $68.80 per 8-GPU instance-hour | source |
| Baseten | $9.98 | Managed endpoint | 180GB | Billed per minute ($0.16633/min) | source |
| Fireworks | $13.00 | Managed endpoint | 180GB | On-demand deployment | source |
| Vast.ai | Live prices | Marketplace | Prices are set by individual hosts and change constantly. Check current offers on Vast.ai. | source |
Per-GPU prices for 8-GPU instances are the instance price divided by 8. Reserved and preemptible prices are lower and are listed in the notes where the provider publishes them. Last checked 2026-10-12.