NVIDIA GH200 Grace Hopper vs NVIDIA H200 SXM5

Side-by-side comparison of NVIDIA GH200 Grace Hopper (96 GB VRAM, 4.0 TB/s + 512 GB/s) and NVIDIA H200 SXM5 (141 GB VRAM, 4.8 TB/s). Compare specs, compute throughput, and workload sizing for LLM inference.

Decision Summary

Spec	NVIDIA GH200 Grace Hopper	NVIDIA H200 SXM5
VRAM	96GB/144GB HBM3 + 480GB LPDDR5X	141GB HBM3e
Memory Bandwidth	4.0 TB/s + 512 GB/s	4.8 TB/s
FP8 TFLOPS	1,979	1,979
FP16 TFLOPS	989	989
Interconnect	NVLink-C2C (900 GB/s)	NVLink 4.0 (900 GB/s)
TDP	1000W	700W
Recommended Quantization	FP8 native — CPU-GPU coherence eliminates explicit offloading	FP8 / FP4 native

Compare for a Workload

Configure a workload to see how each selected GPU performs. Calculations are deterministic estimates based on architectural specifications.

Workload

Context Length

Precision

Batch Size

Configure workload parameters above and click "Calculate VRAM Requirement" to see results.

Cloud Provider Pricing

Current spot and on-demand rates across providers for NVIDIA GH200 Grace Hopper and NVIDIA H200 SXM5.

Featured GPU Pricing Pages

NVIDIA H100 SXM5

80GB HBM3 • 3.35 TB/s • NVLink 4.0

NVIDIA A100 40GB

40GB HBM2 • 1.55 TB/s • NVLink 3.0

NVIDIA A100 80GB

Verified spot rates & reserved pricing

NVIDIA A10G

24GB GDDR6 • 0.62 TB/s • PCIe 4.0

NVIDIA RTX 3090

24GB GDDR6X • 0.94 TB/s • PCIe 4.0

NVIDIA T4

16GB GDDR6 • 0.32 TB/s • PCIe 3.0

Provider	GPU & VRAM	Interconnect	Spot Price ($/hr)	On-Demand ($/hr)	Monthly ($/720h)	Status	Action
Vast.ai Community	H200141GB HBM3e Llama 70B (FP8) ✅DeepSeek R1 ✅Production vLLM Cluster ✅Qwen 32B ✅	NVLink 4.0 (900 GB/s)	$2.79/hr Calculated EstimateMEDIUM SourceVast.ai VerifiedSep 26, 2026 Value 2.79 /hr Methodology No distinct spot listing — the provider's listed hourly (on-demand) rate is surfaced as the tracked rate. Assumptions & Parameters note: Spot rates fluctuate with capacity Refreshed daily from provider APIs and market scrapingSep 26, 2026	$2.79 / hr	$1,707 / mo	Instant	Deploy →
Spheron Bare Metal	H200141GB HBM3e Llama 70B (FP8) ✅DeepSeek R1 ✅Production vLLM Cluster ✅Qwen 32B ✅	NVLink 4.0 (900 GB/s)	$3.19/hr Calculated EstimateMEDIUM SourceSpheron VerifiedSep 26, 2026 Value 3.19 /hr Methodology No distinct spot listing — the provider's listed hourly (on-demand) rate is surfaced as the tracked rate. Assumptions & Parameters note: Spot rates fluctuate with capacity Refreshed daily from provider APIs and market scrapingSep 26, 2026	$3.19 / hr	$1,952 / mo	Instant	Deploy →
Lambda Labs Dedicated	H200141GB HBM3e Llama 70B (FP8) ✅DeepSeek R1 ✅Production vLLM Cluster ✅Qwen 32B ✅	NVLink 4.0 (900 GB/s)	$3.99/hr Calculated EstimateMEDIUM SourceLambda Labs VerifiedSep 26, 2026 Value 3.99 /hr Methodology No distinct spot listing — the provider's listed hourly (on-demand) rate is surfaced as the tracked rate. Assumptions & Parameters note: Spot rates fluctuate with capacity Refreshed daily from provider APIs and market scrapingSep 26, 2026	$3.99 / hr	$2,442 / mo	Instant	Deploy →
RunPod Cloud	H200141GB HBM3e Llama 70B (FP8) ✅DeepSeek R1 ✅Production vLLM Cluster ✅Qwen 32B ✅	NVLink 4.0 (900 GB/s)	$4.31/hr Calculated EstimateMEDIUM SourceRunPod VerifiedSep 26, 2026 Value 4.31 /hr Methodology No distinct spot listing — the provider's listed hourly (on-demand) rate is surfaced as the tracked rate. Assumptions & Parameters note: Spot rates fluctuate with capacity Refreshed daily from provider APIs and market scrapingSep 26, 2026	$4.31 / hr	$2,638 / mo	Instant	Deploy →

On-demand and monthly figures are the providers' listed rates (monthly = listed rate, else hourly × 720h). Rows without a tracked rate show —. All rates subject to preemption and provider availability.

Data Freshness: Public Cloud APIs & Market Scraping | Refreshed Daily (UTC)|Benchmark Baseline: Ubuntu 24.04, CUDA 12.4, vLLM v0.6.x, PagedAttention v2, FlashAttention-3

Prices verified daily from Spheron, RunPod, Vast.ai, and Lambda Labs APIs.

Microarchitecture & Interconnect

NVIDIA GH200 Grace Hopper

ArchitectureGrace Hopper Superchip

Process Node4nm TSMC (GPU) + 5nm TSMC (CPU)

FP8 TFLOPS1,979

FP16 TFLOPS989

Compute BoundMemory-bandwidth bound — 4 TB/s HBM3 + 512 GB/s LPDDR5X

Recommended TopologyMulti-GPU via NVLink-C2C, Grace CPU as memory expander

Recommended Quantization: FP8 native — CPU-GPU coherence eliminates explicit offloading

NVIDIA H200 SXM5

ArchitectureHopper GH200

Process Node4nm TSMC

FP8 TFLOPS1,979

FP16 TFLOPS989

Compute BoundMemory-bandwidth bound across all batch sizes

Recommended Topology8-way HGX Baseboard with NVLink 4.0 Mesh

Recommended Quantization: FP8 / FP4 native

Related GPU Comparisons

Explore canonical side-by-side comparisons for this hardware tier.

NVIDIA GH200 Grace Hopper vs NVIDIA H100 SXM5→

96GB/144GB HBM3 + 480GB LPDDR5X vs 80GB HBM3

NVIDIA B200 Blackwell vs NVIDIA GH200 Grace Hopper→

192GB HBM3e vs 96GB/144GB HBM3 + 480GB LPDDR5X

NVIDIA H100 SXM5 vs NVIDIA H200 SXM5→

80GB HBM3 vs 141GB HBM3e

NVIDIA B200 Blackwell vs NVIDIA H200 SXM5→

192GB HBM3e vs 141GB HBM3e

Deploy NVIDIA GH200 Grace Hopper Deploy NVIDIA H200 SXM5

Data Freshness: Verified via Public Cloud APIs & Market Scraping | Refreshed Daily (UTC)|Benchmark Baseline: Ubuntu 24.04, CUDA 12.4, vLLM v0.6.x (PagedAttention v2, FlashAttention-3), BF16/FP8 weights.

Methodology →