H200 Cloud Rental Pricing
Hopper with 141GB HBM3e, the step up from H100 for large-model inference that needs more memory per GPU.
21 offers from 20 providers · updated November 22, 2026
| Provider | Variant | VRAM | On-demand /hr | Spot /hr | Per month | Billing |
|---|---|---|---|---|---|---|
| Vast.ai | H200 NVL | 141GB | $3.00 | — | $2,190 | per second |
| Vast.ai | H200 | 141GB | $3.76 | — | $2,745 | per second |
| Oblivus | H200 SXM5 | 141GB | $3.99 | — | $2,913 | per hour |
| Hyperstack | H200 SXM | 141GB | $3.99 | — | $2,913 | per minute |
| Crusoe | H200 SXM | 141GB | $4.29 | — | $3,132 | per minute |
| DigitalOcean credits | H200 SXM | 141GB | $4.47 | — | $3,263 | per second |
| Modal credits | H200 SXM | 141GB | $4.54 | — | $3,314 | per second |
| Jarvislabs | H200 SXM | 141GB | $4.59 | $2.29 | $3,351 | per minute |
| Runcrate | H200 | 141GB | $5.13 | — | $3,745 | per second |
| RunPod credits | H200 | 141GB | $5.29 | — | $3,862 | per ms |
| Arvica Cloud | H200 SXM | 141GB | $5.29 | — | $3,862 | per hour |
| Nebius AI Cloud | H200 SXM | 141GB | $5.40 | $0.79 | $3,942 | per second |
| Together AI | H200 | 141GB | $5.99 | $2.99 | $4,373 | per hour |
| Fluence | H200 SXM | 141GB | $6.10 | — | $4,453 | per hour |
| CoreWeave | HGX H200 (min 8) | 141GB | $6.31 | $2.62 | $4,606 | per hour |
| Verda | H200 | 141GB | $6.70 | $3.35 | $4,892 | per hour |
| Prime Intellect | H200 SXM | 141GB | $6.75 | $3.41 | $4,928 | per minute |
| AWS credits | H200 SXM (min 8) | 141GB | $7.91 | $4.82 | $5,776 | per second |
| Oracle Cloud | H200 SXM (min 8) | 141GB | $10.00 | — | $7,300 | per second |
| Microsoft Azure | H200 SXM (min 8) | 141GB | $10.60 | — | $7,738 | per second |
| Google Cloud | H200 SXM (min 8) | 141GB | $10.60 | $6.36 | $7,739 | per second |
Per GPU-hour, on-demand unless noted. Monthly = 730 hours. Check each provider for region availability and multi-GPU node minimums.
H200 guides
H100 Rental Prices in 2026: The Cheapest Clouds per GPU-Hour
H100 rental prices from $1.67 per GPU-hour. Compare marketplaces, GPU clouds, AWS, Google Cloud and Azure, plus monthly, 8-GPU node, A100, H200 and B200 costs.
Nov 2026Together AI Pricing in 2026: Serverless, Dedicated Endpoints and Fine-Tuning
Together AI API pricing: serverless from $0.15/1M tokens, Kimi K3 $3/$15, H100 clusters $3.99/hr, dedicated $5.49/hr, fine-tuning from $0.34/1M. No free trial.
H200 rental FAQ
How much does it cost to rent an H200?
On-demand H200 rental ranges from $3.00 to $10.60 per GPU-hour across 20 providers (median $5.29), as of November 22, 2026. Spot/interruptible capacity starts at $0.79/hr.
What is the cheapest H200 cloud provider?
Vast.ai has the lowest on-demand H200 price we track at $3.00 per GPU-hour (H200 NVL).
How much does an 8x H200 server cost per month?
At the cheapest on-demand rate ($3.00/GPU-hour), 8x H200 running 24/7 costs about $17,520 per month (730 hours). At the median rate it is about $30,894.