LambdaIn stock
On-Demand from $2.29
From
$2.29 / GPU / hr
On-DemandVisit website →
Rental price index
NVIDIA GH200
🏢Hopper GPU fused with a Grace CPU and shared memory.
Prices collected 2026-09-12 · 3 providers, 3 configurations
Compare vs other GPUs →On-demandOn-demand (archive, Lambda only)
Index median price per GPU per hour, by billing type. Nothing is interpolated.
At a glance
Median on-demand
$3.02/GPU/hr
Cheapest on-demand
$2.29Lambda
LambdaFloor, any billing
$2.29incl. spot
Per month at 720h
$2,174
NVIDIA GH200 pricing by provider
One card per provider, cheapest first. A configuration is one priced unit — an instance type, a SKU, a marketplace offer — and the per-GPU price is that unit's price divided by its GPU count. Open a card to see every configuration with its vCPU, RAM and region where the provider publishes them.
3 providers · 3 configurations
SpheronIn stock
1 config1xcollected 2026-09-12
On-Demand from $3.02
From
$3.02 / GPU / hr
On-DemandVisit website →
CoreWeaveIn stock
USUSA1 config1xcollected 2026-09-12
On-Demand from $6.50
From
$6.50 / GPU / hr
On-DemandVisit website →What fits on a NVIDIA GH200
Weights + KV cache at 8K context + runtime overhead, against 86 GB usable of the 96 GB on the card. Capacity follows total parameters, even for mixture-of-experts models — any expert may be needed next. Where one card is not enough, the table says how many are, and prices the set at today's median.
| Model | Params | 4-bit | 8-bit | 16-bit |
|---|---|---|---|---|
| 2.8T104B active | 19× GPU$57.38/hr | 35× GPU$105.70/hr | 65× GPU$196.30/hr | |
| 552B16B active | 4× GPU$12.08/hr | 7× GPU$21.14/hr | 13× GPU$39.26/hr | |
 Llama 4 Maverick | 400B17B active | 3× GPU$9.06/hr | 5× GPU$15.10/hr | 10× GPU$30.20/hr |
 gpt-oss-120b | 117B5.1B active | ✓ 67.9 GB~1115 tok/s | 2× GPU$6.04/hr | 3× GPU$9.06/hr |
| Llama 3.1 70B | 70B | ✓ 44.5 GB~81 tok/s | ✓ 79.5 GB~43 tok/s | 2× GPU$6.04/hr |
 Gemma 4 31B | 31B | ✓ 21.2 GB~184 tok/s | ✓ 36.7 GB~97 tok/s | ✓ 65.7 GB~52 tok/s |
 Qwen3 Coder 30B-A3B | 30.5B3.3B active | ✓ 19 GB~1724 tok/s | ✓ 34.3 GB~913 tok/s | ✓ 62.9 GB~485 tok/s |
| Qwen3.8 27B | 27B | ✓ 18.7 GB~211 tok/s | ✓ 32.2 GB~112 tok/s | ✓ 57.6 GB~59 tok/s |
| gpt-oss-20b | 21B3.6B active | ✓ 13.7 GB~1580 tok/s | ✓ 24.2 GB~837 tok/s | ✓ 43.9 GB~444 tok/s |
| Llama 3.1 8B | 8B | ✓ 6.9 GB~711 tok/s | ✓ 10.9 GB~376 tok/s | ✓ 18.4 GB~200 tok/s |
tok/s is decode throughput from memory bandwidth, derated 20%, single stream — an estimate, not a benchmark. Set cost uses today's median × GPUs required.
Specifications
Memory96 GB HBM3
Memory bandwidth4,000 GB/s
ArchitectureGrace Hopper
VendorNVIDIA
ClassDatacenter
Released2023-08
Bandwidth is listed because it governs generation speed: each token streams the active weights once, so tokens per second is bandwidth divided by bytes read. TFLOPs decide prompt processing and training, not how fast text appears.
Alternatives
NVIDIA H100 NVL94 GB · $2.95/hr median · 6 providersNVIDIA H100 SXM80 GB · $3.49/hr median · 23 providersNVIDIA A100 80GB80 GB · $1.79/hr median · 19 providers
Questions
How much does it cost to rent a NVIDIA GH200?
As of 2026-09-12, the median on-demand price across 3 providers is $3.02 per GPU-hour, about $2,174 a month at 720 hours. The cheapest we saw was $2.29 at Lambda.
Why do prices for the same GPU vary so much?
Hyperscalers bundle CPU, RAM and network into an instance price and charge for the ecosystem around it. Neoclouds sell the GPU more directly. Marketplaces are independent hosts competing on price, with the trade-offs of shared, third-party hardware. The per-GPU figure is the instance price divided by GPU count, which is the only way to compare an 8-GPU node with a single-card listing.
Is the median the price I will pay?
No — it is the middle of the market. Each provider's own median is taken first so a provider with many instance sizes counts once, then the median across providers. Spot and reserved rates are excluded from it; they set the floor shown separately. Use it to judge whether a quote is high or low, then verify with the provider.
Which models can a NVIDIA GH200 run?
With 96 GB per GPU, the table above shows which reference models fit on one card at 4, 8 and 16-bit, using the same memory arithmetic as the rest of Runyard: weights plus KV cache plus runtime overhead, with capacity following total parameters even for mixture-of-experts models. Where a model needs more than one GPU it says how many and what that set costs at today's median.
Own the hardware instead? Work out what your card holds with the same arithmetic.
Check what fits →