NVIDIA GH200 96GB HBM3 vs NVIDIA H100 PCIe 80GB

The GH200 delivers 31% more FP8 throughput than the H100 (1,979 vs 1,513 TFLOPS dense). The GH200 also carries 16 GB more memory (96 GB vs 80 GB).

GH200: Hopper, 2023 H100: Hopper, 2022

FP8 dense lead

+31%

GH200: 1,979 vs 1,513 TFLOPS

Memory

96 vs 80 GB

16 GB more for the GH200

Bandwidth

4.0 TB/s vs 2.0 TB/s

GH200 moves data faster

TDP

1.0 kW vs 350 W

H100 draws 650 W less

See rental pricing from $1.66/hr Download comparison sheet Printable version →

NVIDIA GH200 96GB HBM3

Leads on FP8 Full specs →

What is the NVIDIA GH200?

The GH200 is a GPU from NVIDIA on the Hopper architecture, an accelerator for AI training, inference and high-performance computing. It is a superchip that pairs a CPU and accelerators on one board for rack-scale systems. GH200 full specifications →

Can I rent the GH200 in the cloud?

Yes. Cloud providers tracked by Flopper rent the GH200 by the hour, and the pricing page compares their current rates and the price history side by side. GH200 rental prices →

How does the GH200 compare with other accelerators?

This page sets the GH200 against the H100. The GH200 is the newer generation of the two. Both are NVIDIA parts; the GH200 is a superchip while the H100 is a PCIe card. Flopper also compares it head to head with the H100 SXM5 80GB, across every published precision, memory and power. GH200 vs H100 SXM5 80GB →

Where do the GH200 figures come from?

Every GH200 figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

NVIDIA H100 PCIe 80GB

Full specs →

What is the NVIDIA H100?

The H100 is a GPU from NVIDIA on the Hopper architecture, an accelerator for AI training, inference and high-performance computing. It is a PCIe add-in card, so it fits standard servers and workstations. H100 full specifications →

Can I rent the H100 in the cloud?

Yes. Cloud providers tracked by Flopper rent the H100 by the hour, and the pricing page compares their current rates and the price history side by side. H100 rental prices →

How does the H100 compare with other accelerators?

This page sets the H100 against the GH200. The H100 is the older generation of the two. Both are NVIDIA parts; the H100 is a PCIe card while the GH200 is a superchip. Flopper also compares it head to head with the H100 SXM5 80GB, across every published precision, memory and power. H100 vs H100 SXM5 80GB →

Where do the H100 figures come from?

Every H100 figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

Specifications

GH200

H100

Architecture

HopperHopper

Launch Year

20232022

Form Factor

SuperchipPCIe

Memory

96 GB80 GB

Memory Bandwidth

4.0 TB/s2.0 TB/s

TDP

1.0 kW350 W

Process Node

4nm4nm

Spec

GH200

H100

Architecture HopperHopper

Launch Year 20232022

Form Factor SuperchipPCIe

Memory 96 GB80 GB

Memory Bandwidth 4.0 TB/s2.0 TB/s

TDP 1.0 kW350 W

Process Node 4nm4nm

Performance (TFLOPS)

GH200

H100

34 TFLOPS 26 TFLOPS

67 TFLOPS 51 TFLOPS

494 TFLOPS

989 TFLOPS sparse

378 TFLOPS

sparse not published

990 TFLOPS

1,979 TFLOPS sparse

757 TFLOPS

sparse not published

990 TFLOPS

1,979 TFLOPS sparse

757 TFLOPS

sparse not published

1,979 TFLOPS

3,958 TFLOPS sparse

1,513 TFLOPS

sparse not published

FP6

1,979 TOPS

3,958 TOPS sparse

1,513 TOPS

sparse not published

Tensor Core

67 TFLOPS

Precision

GH200

H100

34 TFLOPS 26 TFLOPS

67 TFLOPS 51 TFLOPS

494 TFLOPS

989 TFLOPS sparse

378 TFLOPS

sparse not published

990 TFLOPS

1,979 TFLOPS sparse

757 TFLOPS

sparse not published

990 TFLOPS

1,979 TFLOPS sparse

757 TFLOPS

sparse not published

1,979 TFLOPS

3,958 TFLOPS sparse

1,513 TFLOPS

sparse not published

FP6

1,979 TOPS

3,958 TOPS sparse

1,513 TOPS

sparse not published

Tensor Core

67 TFLOPS

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

The GH200 is the newer part: Hopper (2023) against the H100's Hopper (2022).

The GH200 carries 16 GB more memory (96 GB against 80 GB), so larger models fit on one GPU before you shard. Bandwidth is 4.0 TB/s against 2.0 TB/s, which sets the ceiling on decode speed.

Power is 1.0 kW against 350 W, and at FP32 the H100 does more work per watt. Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP64, FP32, TF32, BF16, FP16, FP8, INT8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

Is the GH200 faster than the H100?

At FP8 precision the GH200 reaches 1,979 TFLOPS dense against 1,513 TFLOPS for the H100. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the GH200 or the H100?

Choose the GH200 for training and high-throughput inference: it has 31% more FP8 throughput than the H100. It also carries 16 GB more memory, so larger models fit on one GPU before you shard. The GH200 draws 650 W more power, so budget for cooling and power delivery before you commit. Packaging differs too: the GH200 is a superchip while the H100 is a PCIe card, so check what your servers accept.

Which has more memory, the GH200 or the H100?

The GH200 carries 96 GB of memory versus 80 GB for the H100. Memory bandwidth is 4.0 TB/s for the GH200 and 2.0 TB/s for the H100.

How much power do the GH200 and the H100 draw?

The GH200 is rated at 1.0 kW TDP and the H100 at 350 W. On FP32 throughput per watt, the H100 is the more efficient part.

How much do the GH200 and the H100 cost to rent?

Live cloud listings tracked by Flopper price the GH200 from $2.29 per GPU-hour across 4 providers, and the H100 from $1.66 per GPU-hour across 12 providers. The rental section on this page links to every provider, rate and the price history for each GPU.

Where to Rent

NVIDIA GH200 96GB HBM3

Rent from

$2.29 / GPU-hr

Lowest of 4 live listings across 4 providers · median $3.63

See all GH200 rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

NVIDIA H100 PCIe 80GB

Rent from

$1.66 / GPU-hr

Lowest of 21 live listings across 12 providers · median $2.70

See all H100 rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

Open in Full Compare Tool → GH200 full specs H100 full specs Estimate cluster cost in the calculator →

NVIDIA GH200 96GB HBM3 vs NVIDIA H100 PCIe 80GB

NVIDIA GH200 96GB HBM3

What is the NVIDIA GH200?

Can I rent the GH200 in the cloud?

How does the GH200 compare with other accelerators?

Where do the GH200 figures come from?

NVIDIA H100 PCIe 80GB

What is the NVIDIA H100?

Can I rent the H100 in the cloud?

How does the H100 compare with other accelerators?

Where do the H100 figures come from?

Specifications

Performance (TFLOPS)

FLOPS by Precision

What actually differs

Frequently asked questions

Is the GH200 faster than the H100?

Which should you choose, the GH200 or the H100?

Which has more memory, the GH200 or the H100?

How much power do the GH200 and the H100 draw?

How much do the GH200 and the H100 cost to rent?

Where to Rent

NVIDIA GH200 96GB HBM3

NVIDIA H100 PCIe 80GB

Related comparisons

Get Comparison Updates

Stay Updated on GPU Specs