H200 NVL rental prices: cheapest $/hr today
The cheapest on-demand H200 NVL rental today is $3.62/hr on Massed Compute, updated 2026-10-02. 143 GB VRAM, NVIDIA.
Pricing at a glance
All current H200 NVL offers
2 offers. Rates are per GPU; total is for the smallest listed node.
| GPU / provider | Billing | $/GPU-hour | GPUs / total per hour | Availability | CPU / RAM / disk | Region | Updated | Action |
|---|---|---|---|---|---|---|---|---|
| H200 NVL Massed Compute | on-demand | $3.62+0.0% over 30 days | 1× $3.62 | Not reported | vCPU not reportedRAM not reportedDisk not reported | Not reported | 2026-10-02 | View providerReport this price |
| H200 NVL RunPod | secure cloud | $3.79+0.0% over 30 days | Node minimum and total not reported | Not reported | vCPU not reportedRAM not reportedDisk not reported | Not reported | 2026-10-02 | View providerReport this price |
Tables rank by price, lowest first. Commercial relationships never affect ordering.How we collect and rank prices.
On-demand and secure rates do not require a reservation. Community and spot capacity can be interrupted. Prices are per GPU per hour, before storage/egress. Every row shows its update date.
H200 NVL hardware specs
| Architecture | Hopper |
|---|---|
| VRAM | 143 GB |
| FP16 TFLOPS (dense) | 835 |
| Memory bandwidth | 4800 GB/s |
| TDP | 600 W |
| Interconnect | NVLink 900GB/s |
| Launched | 2024 |
Datasheet spec for the H200 NVL variant. Source: www.pny.com.
Compare H200 NVL configurations
H200 NVL purchase prices · Rent or buy calculator · Compare two GPUs
LLM memory estimates for this GPU
These models have an estimated inference footprint within 143 GB at one or more precisions. The scenario uses 4,096 tokens, batch 1, a 16-bit KV cache and 20% runtime headroom. A memory fit does not establish speed or software compatibility.
| Model | 16-bit GB | 8-bit GB | 4-bit GB |
|---|---|---|---|
| Qwen2.5 0.5B | 1.2 ✓ | 0.7 ✓ | 0.4 ✓ |
| Qwen2.5 1.5B | 3.8 ✓ | 2.0 ✓ | 1.1 ✓ |
| Qwen2.5 3B | 7.6 ✓ | 3.9 ✓ | 2.0 ✓ |
| Qwen2.5 7B | 18.6 ✓ | 9.4 ✓ | 4.9 ✓ |
| Qwen2.5 14B | 36.4 ✓ | 18.7 ✓ | 9.8 ✓ |
| Qwen2.5 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| Qwen2.5 72B | 176.1 | 88.9 ✓ | 45.2 ✓ |
| Qwen2.5 Coder 0.5B | 1.2 ✓ | 0.7 ✓ | 0.4 ✓ |
| Qwen2.5 Coder 1.5B | 3.8 ✓ | 2.0 ✓ | 1.1 ✓ |
| Qwen2.5 Coder 3B | 7.6 ✓ | 3.9 ✓ | 2.0 ✓ |
| Qwen2.5 Coder 7B | 18.6 ✓ | 9.4 ✓ | 4.9 ✓ |
| Qwen2.5 Coder 14B | 36.4 ✓ | 18.7 ✓ | 9.8 ✓ |
| Qwen2.5 Coder 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| Qwen3 0.6B | 2.4 ✓ | 1.5 ✓ | 1.0 ✓ |
| Qwen3 1.7B | 5.4 ✓ | 3.0 ✓ | 1.8 ✓ |
| Qwen3 4B | 10.4 ✓ | 5.6 ✓ | 3.1 ✓ |
| Qwen3 8B | 20.4 ✓ | 10.6 ✓ | 5.6 ✓ |
| Qwen3 14B | 36.2 ✓ | 18.5 ✓ | 9.7 ✓ |
| Qwen3 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| Qwen3 30B A3B | 73.8 ✓ | 37.1 ✓ | 18.8 ✓ |
| Qwen3 235B A22B | 565.2 | 283.1 | 142.0 ✓ |
| DeepSeek R1 Distill Qwen 1.5B | 4.4 ✓ | 2.3 ✓ | 1.2 ✓ |
| DeepSeek R1 Distill Qwen 7B | 18.6 ✓ | 9.4 ✓ | 4.9 ✓ |
| DeepSeek R1 Distill Qwen 14B | 36.4 ✓ | 18.7 ✓ | 9.8 ✓ |
| DeepSeek R1 Distill Qwen 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| DeepSeek R1 Distill Llama 8B | 19.9 ✓ | 10.3 ✓ | 5.5 ✓ |
| DeepSeek R1 Distill Llama 70B | 170.9 | 86.3 ✓ | 43.9 ✓ |
| Mistral 7B v0.3 | 18.0 ✓ | 9.3 ✓ | 5.0 ✓ |
| Mixtral 8x7B v0.1 | 112.7 ✓ | 56.7 ✓ | 28.7 ✓ |
| Mixtral 8x22B v0.1 | 338.6 | 169.9 | 85.5 ✓ |
| Mistral Small 24B 2501 | 57.4 ✓ | 29.1 ✓ | 14.9 ✓ |
| Mistral Nemo 2407 | 30.2 ✓ | 15.5 ✓ | 8.2 ✓ |
| phi 4 | 36.2 ✓ | 18.6 ✓ | 9.8 ✓ |
Compare model memory requirements or change context and memory assumptions on a model's page.
When to choose H200 NVL
Use the model memory estimates above to check whether your weights, context and batch size fit within 143 GB. Then compare the exact variant, node size and software support. Capacity alone does not establish inference speed or training throughput.
If your workload fits within 96 GB, compare RTX PRO 6000 rental prices from $1.10/GPU-hour. The smaller memory budget is the trade-off; price alone does not make the cards interchangeable.
H200 NVL price changes
- 30 days: 0.0% across 2 matching firm-price configurations since 2026-09-02.
- 90 days: No matching dated observations for this comparison.
Changes use the geometric mean of price ratios for the same provider, variant, tier, GPU count, instance and region. New listings are not counted as price changes.
Price history
on-demand rate per day, since 2026-07-06: $3.62 low · $3.79 high. Full ledger on the history hub.
FAQ
The cheapest H200 NVL rental we have is $3.62 per GPU-hour on Massed Compute (on-demand), updated 2026-10-02.
143 GB.
Massed Compute has the cheapest verified on-demand rate today at $3.62 per GPU-hour. Verified 2026-10-02.
Related
All 140 GB+ GPUs · GPUs under $2/hr · Price movers this week · Massed Compute review
Should you rent or buy?
Compare H200 NVL purchase prices and quote availability, then enter your usage and electricity costs in the rent-versus-buy calculator. Purchase listings have their own variants, stock status and update dates.