H200 SXM

Every provider's rate, side by side.

H200 SXM is available through the OneInfer GPU marketplace with live provider pricing, regions, VRAM, and deployment options from the API.

7 providers list it in 18 regions, from $3.50 per GPU-hour with nebius_ai. It has 141 GB of GPU memory. Pick a provider and region, deploy it from the OneInfer Console with your own Docker image, connect over SSH, and pay from one pay-as-you-go wallet while it runs.

Deploy the H200 SXM in 4 steps

Step 1 of 4

7 providers list the H200 SXM, from $3.50 per GPU-hour. The table below sorts them by price.

Cloud GPU
VRAM
141GB
CUDA cores
16,896
TDP
700W
Bandwidth
N/A
Process node
5nm
Launch date
N/A

Where you can run it

The H200 SXM is available now in 18 of the 94 regions OneInfer has GPUs in. Pick a region to see who offers it there, or see every GPU by region.

Available nowNot available

Americas

Europe

Asia-Pacific

Prices are per GPU-hour, the lowest in each region. Grey regions have other GPUs, not the H200 SXM.

Available from 7 providers

Sorted by price. Select the price heading to re-sort.

H200 SXM providers, regions, configurations and hourly prices
ProviderRegionsConfigAvailabilityDeploy
nebius_ainebius_ai
EU-FI-011x 141GB$3.50Available
nebius_ainebius_ai
US-MO-018x 141GB$3.50$28.00/hr for 8 GPUsAvailable
vastaivastai
CA-BC-01, AS-JP-01, EU-RO-012x 141GB$3.62$7.24/hr for 2 GPUsAvailable
vastaivastai
CA-BC-01, US-GA-01, AS-JP-01, CA-QC-01, EU-RO-011x 141GB$3.63Available
runpodrunpod
US-CA-02, EU-IS-041x 141GB$5.29Available
aws
AS-KR-01, CA-QC-01, IN-MH-01, JP-TYO-01, SE-AB-01, US-CA-01, US-OH-01, US-OR-01, US-VA-018x 141GB$7.91$63.30/hr for 8 GPUsAvailable
aws
GB-ENG-01, OC-AU-NSW-01, US-OH-01, US-OR-018x 141GB$7.91$63.30/hr for 8 GPUsAvailable

Full specifications

Data from GPU specs and provider APIs

AI TOPS
3,958
FP8 TFLOPS
3,958
FP16 TFLOPS
1,979
BF16 TFLOPS
1,979
FP32 TFLOPS
67
INT8 TOPS
3,958
Sparse multiplier
2

Benchmarks

H200 SXM vs comparable GPUs on the platform

Inference throughput

7B model, tokens per second

B200 SXM15,000
H200 SXM8,500
H100620
A100500
GeForce RTX 5090180

Training performance

Overall training multiplier

B200 SXM30x
H200 SXM18x
H10015x
A1003.4x
L40S2.8x

GeekBench OpenCL

Compute benchmark score

GeForce RTX 5090367,740
H200 SXM343,598
L40S337,706
H100336,474
B200 SXM333,768

Ready to deploy H200 SXM?

Deploy from your choice of 7 providers.