Hopper GPU

NVIDIA H200 SXM 141GB

141GB HBM3e | 989.5 TFLOPS FP16 | 4.8 TB/s bandwidth

From $3.49 /hr

LLM training Large model inference HPC Generative AI

Specifications

VRAM

141 GB HBM3e

FP16

989.5 TFLOPS

FP8

2.0 PFLOPS

FP4

N/A

Memory BW

4.8 TB/s

TDP

700W

Interconnect

NVLink 4

Architecture

Hopper

Cloud GPU Pricing (4 offers)

ProviderInstance TypevCPUsRAMPrice/hrPrice/moSpot PriceAvailabilityAction
Lambda CloudCheapest
gpu_1x_h20048480 GB$3.49/hr$2,547.70/mo--AvailableDeploy on Lambda
CoreWeave
h200-sxm-1x48480 GB$3.85/hr$2,810.50/mo--AvailableDeploy on CoreWeave
Google Cloud Platform
a3-ultragpu-1g96680 GB$4.20/hr$3,066.00/mo
$2.52-40%
AvailableDeploy on Google Cloud
Amazon Web Services
p5e.24xlarge96768 GB$4.50/hr$3,285.00/mo
$2.70-40%
AvailableDeploy on AWS
Last updated 57h ago

Compare With Other GPUs

Frequently Asked Questions

What is the cheapest NVIDIA H200 SXM 141GB cloud provider?
The cheapest NVIDIA H200 SXM 141GB is available on Lambda Cloud at $3.49/hr (gpu_1x_h200).
How much does NVIDIA H200 SXM 141GB cost per hour?
NVIDIA H200 SXM 141GB cloud GPU pricing ranges from $3.49/hr to $4.50/hr depending on the provider and configuration.
What are the specs of NVIDIA H200 SXM 141GB?
NVIDIA H200 SXM 141GB features 141GB HBM3e memory, 989.5 TFLOPS FP16 performance, 4.8 TB/s memory bandwidth, and 700W TDP. Architecture: Hopper.
Is NVIDIA H200 SXM 141GB good for AI training?
Yes, NVIDIA H200 SXM 141GB is well-suited for AI training workloads. Key use cases include: LLM training, Large model inference, HPC, Generative AI.

Related GPUs