H200 NVL 141GB, what it runs and what it costs to rent
Specs checked · Rental prices from · Aliteq Team · How we compute this
Some links on this page are referral links: we may earn a commission, at no cost to you.
The short answer
The NVIDIA H200 NVL has 141 GB of VRAM and 4800 GB/s of memory bandwidth (2024). At Q4_K_M quality or better with an 8K context it runs 53 of the open models we track; the largest that fits with room to spare is Qwen3.5 122B-A10B (MoE). The cheapest rental right now is $3.000/hr (Vast.ai, on demand).
- VRAM
- 141 GB
- Bandwidth
- 4800 GB/s
- Board power
- 600 W
- Launched
- 2024
Source: NVIDIA specification · bandwidth not yet confirmed by a second source, so we don't estimate speed from it
Rent the H200 NVL
| Provider | Type | From, per hour | Rent |
|---|---|---|---|
| Vast.ai | On demand | $3.000/hr | Vast.ai |
| Lium | On demand | $3.600/hr | Lium |
| Runpod | On demand | $3.790/hr | Runpod |
Lowest price per provider in our hourly snapshot (). Rent buttons are referral links. Every card side by side: cloud GPU prices.
Before you rent the H200 NVL: what an H200 rental costs · Runpod vs Vast.ai vs Lambda · is Vast.ai safe · is Runpod worth it · rent or buy · what an H200 costs to buy · every AI GPU's buy and rent price.
H200 NVL rental price history
The cheapest H200 NVL offer we saw each day on the providers we track, since Aug 20. Renting it on demand costs $3.00/hr now; the lowest since Nov 2 was $1.87/hr on Nov 3 (Vast.ai), so today is 61% above that.
- On demand now
- $3.00/hr
- Lowest on demand since Nov 2
- $1.87/hr · Nov 3
- Cheapest offer now
- $3.00/hr
- Lowest offer since Aug 20
- $0.400/hr · Nov 1
- Cheapest offer, incl. interruptible
- Cheapest on demand (from Nov 2)
Daily minimum of our hourly readings of Vast.ai, Runpod and Lium. Interruptible (spot) offers can be taken back mid-job. Before Nov 2 Vast mixed spot into on-demand and Runpod listed Community Cloud rates, so the on-demand line starts then. Every number is in our open daily price file.
Every day as a table
| Day | Cheapest offer | Cheapest on demand |
|---|---|---|
| Nov 18 (now) | $3.00/hr · Vast.ai | $3.00/hr · Vast.ai |
| Nov 17 | $0.667/hr · Vast.ai | $2.99/hr · Lium |
| Nov 16 | $1.00/hr · Vast.ai | $2.99/hr · Lium |
| Nov 15 | $1.32/hr · Vast.ai | $3.27/hr · Vast.ai |
| Nov 14 | $1.32/hr · Vast.ai | $3.27/hr · Vast.ai |
| Nov 13 | $1.32/hr · Vast.ai | $2.99/hr · Vast.ai |
| Nov 12 | $1.32/hr · Vast.ai | $3.27/hr · Vast.ai |
| Nov 11 | $0.402/hr · Vast.ai | $3.27/hr · Vast.ai |
| Nov 10 | $0.402/hr · Vast.ai | $3.27/hr · Vast.ai |
| Nov 9 | $0.402/hr · Vast.ai | $3.00/hr · Vast.ai |
| Nov 8 | $0.402/hr · Vast.ai | $2.93/hr · Vast.ai |
| Nov 7 | $0.402/hr · Vast.ai | $2.67/hr · Vast.ai |
| Nov 6 | $0.402/hr · Vast.ai | $2.67/hr · Vast.ai |
| Nov 5 | $0.400/hr · Vast.ai | $1.87/hr · Vast.ai |
| Nov 4 | $0.400/hr · Vast.ai | $1.87/hr · Vast.ai |
| Nov 3 | $0.400/hr · Vast.ai | $1.87/hr · Vast.ai |
| Nov 2 | $1.32/hr · Vast.ai | $2.67/hr · Vast.ai |
| Nov 1 | $0.400/hr · Vast.ai | — |
| Oct 31 | $2.67/hr · Vast.ai | — |
| Oct 30 | $2.67/hr · Vast.ai | — |
| Oct 29 | $3.00/hr · Vast.ai | — |
| Oct 28 | $1.87/hr · Vast.ai | — |
| Oct 27 | $2.44/hr · Vast.ai | — |
| Oct 26 | $2.67/hr · Vast.ai | — |
| Oct 25 | $2.89/hr · Vast.ai | — |
| Oct 24 | $2.67/hr · Vast.ai | — |
| Oct 23 | $3.20/hr · Vast.ai | — |
| Oct 22 | $3.20/hr · Vast.ai | — |
| Oct 21 | $3.07/hr · Vast.ai | — |
| Oct 20 | $1.87/hr · Vast.ai | — |
| Oct 19 | $1.87/hr · Vast.ai | — |
| Oct 18 | $1.87/hr · Vast.ai | — |
| Oct 17 | $3.27/hr · Vast.ai | — |
| Oct 16 | $1.61/hr · Vast.ai | — |
| Oct 15 | $3.27/hr · Vast.ai | — |
| Oct 14 | $3.79/hr · Runpod | — |
| Oct 13 | $3.60/hr · Vast.ai | — |
| Oct 12 | $3.71/hr · Vast.ai | — |
| Oct 11 | $3.61/hr · Vast.ai | — |
| Oct 10 | $3.61/hr · Vast.ai | — |
| Oct 9 | $3.67/hr · Vast.ai | — |
| Oct 8 | $3.48/hr · Vast.ai | — |
| Oct 7 | $3.20/hr · Vast.ai | — |
| Oct 6 | $2.67/hr · Vast.ai | — |
| Oct 5 | $2.67/hr · Vast.ai | — |
| Oct 4 | $3.34/hr · Vast.ai | — |
| Oct 3 | $2.67/hr · Vast.ai | — |
| Oct 2 | $2.20/hr · Vast.ai | — |
| Oct 1 | $2.20/hr · Vast.ai | — |
| Sep 30 | $2.07/hr · Vast.ai | — |
| Sep 29 | $2.20/hr · Vast.ai | — |
| Sep 28 | $2.20/hr · Vast.ai | — |
| Sep 27 | $2.59/hr · Vast.ai | — |
| Sep 26 | $2.33/hr · Vast.ai | — |
| Sep 25 | $2.33/hr · Vast.ai | — |
| Sep 24 | $2.59/hr · Vast.ai | — |
| Sep 23 | $2.60/hr · Vast.ai | — |
| Sep 22 | $2.60/hr · Vast.ai | — |
| Sep 21 | $2.60/hr · Vast.ai | — |
| Sep 20 | $2.60/hr · Vast.ai | — |
| Sep 19 | $2.60/hr · Vast.ai | — |
| Sep 18 | $2.60/hr · Vast.ai | — |
| Sep 17 | $2.60/hr · Vast.ai | — |
| Sep 16 | $2.60/hr · Vast.ai | — |
| Sep 15 | $2.60/hr · Vast.ai | — |
| Sep 14 | $2.60/hr · Vast.ai | — |
| Sep 13 | $2.60/hr · Vast.ai | — |
| Sep 12 | $2.60/hr · Vast.ai | — |
| Sep 11 | $2.60/hr · Vast.ai | — |
| Sep 10 | $2.60/hr · Vast.ai | — |
| Sep 9 | $2.60/hr · Vast.ai | — |
| Sep 8 | $2.60/hr · Vast.ai | — |
| Sep 7 | $1.62/hr · Vast.ai | — |
| Sep 6 | $3.34/hr · Vast.ai | — |
| Sep 5 | $3.34/hr · Vast.ai | — |
| Sep 4 | $3.34/hr · Vast.ai | — |
| Sep 3 | $3.34/hr · Vast.ai | — |
| Sep 2 | $3.34/hr · Vast.ai | — |
| Sep 1 | $2.74/hr · Vast.ai | — |
| Aug 31 | $2.74/hr · Vast.ai | — |
| Aug 30 | $2.74/hr · Vast.ai | — |
| Aug 29 | $2.07/hr · Vast.ai | — |
| Aug 28 | $1.94/hr · Vast.ai | — |
| Aug 27 | $2.60/hr · Vast.ai | — |
| Aug 26 | $2.60/hr · Vast.ai | — |
| Aug 25 | $2.60/hr · Vast.ai | — |
| Aug 24 | $2.20/hr · Vast.ai | — |
| Aug 23 | $3.13/hr · Vast.ai | — |
| Aug 22 | $3.12/hr · Vast.ai | — |
| Aug 21 | $1.94/hr · Vast.ai | — |
| Aug 20 | $3.16/hr · Vast.ai | — |
Which AI models run on the H200 NVL
Computed by our VRAM engine for an 8K context: the best quality each model fits at in 141 GB. Speeds are upper-bound estimates from memory bandwidth, not benchmarks.
| Model | Best fit | VRAM used | ~tok/s |
|---|---|---|---|
| Qwen3 235B-A22B (MoE) | Q4_K_M | 135 GB96% · tight | — |
| MiniMax-M2.5 (MoE) | Q4_K_M | 132 GB94% · tight | — |
| Qwen3.8-Flash-Next (MoE) | Q6_K | 138 GB98% · tight | — |
| Qwen3.5 122B-A10B (MoE) | Q8_0 | 125 GB88% | — |
| NVIDIA Nemotron 3.5 Super VL 120B-A12B (MoE) | Q8_0 | 124 GB88% | — |
| NVIDIA Nemotron 3 Super 120B-A12B (MoE) | Q8_0 | 123 GB87% | — |
| Mistral Small 4 119B (MoE) | FP8 | 114 GB81% | — |
| gpt-oss 120B (MoE) | MXFP4 | 62 GB44% | — |
| Llama 4 Scout 17B-16E (MoE) | Q8_0 | 109 GB77% | — |
| GLM-4.5-Air (MoE) | Q8_0 | 107 GB76% | — |
| Qwen3-Next 80B-A3B (MoE) | Q8_0 | 80 GB57% | — |
| Qwen3-Coder-Next 80B-A3B (MoE) | Q8_0 | 80 GB57% | — |
Too big for 141 GB (rent a bigger card instead): what it costs to run MiniMax-M3 (MoE), what it costs to run MiniMax-M3.1-Flash (MoE), what it costs to run GLM-5.3 (MoE), what it costs to run DeepSeek V4.1 Flash (MoE).
Check a model and context of your own in the VRAM calculator.
AI video on the H200 NVL
Alibaba's Wan 2.2 is an open video model you can run yourself in ComfyUI. With 141 GB the H200 NVL fits:
- Wan 2.2 14B at 720p without offloading (80 GB and up, per Runpod's Wan 2.2 guide)
- Wan 2.2 14B in fp16, or in fp8 at 720p (48 GB and up, per Runpod's Wan 2.2 guide)
- Wan 2.2 14B at 640x640 with fp8 weights (24 GB and up, per Runpod's Wan 2.2 guide)
- Wan 2.2 5B, with ComfyUI's native offloading (8 GB and up, per ComfyUI's Wan 2.2 docs)
What each setup needs, the cheapest card to rent and what one clip costs: cloud GPUs for ComfyUI with Flux and Wan 2.2. Against an app or an API: AI video cost per clip, app vs API vs rented GPU. Which open model to run, and the licence catch for US users: the best open source AI video generator. Picking by job instead (YouTube, TikTok, ads, animation)? See the best AI video generator for each job. Rather not run a model yourself? Hosted video models such as Seedance, Veo and Hailuo are compared in Kling AI alternatives, priced per clip.
Our guides on the H200 NVL
Cards to compare it with
- H200 SXM · 141 GB
- RTX PRO 6000 Blackwell Server Edition · 96 GB
- H100 NVL · 94 GB
- AMD Instinct MI300X · 192 GB
- B200 · 192 GB
- H100 SXM · 80 GB
Every card by VRAM tier: all GPUs for gaming and local AI.