Flopper.io Flopper.io

NVIDIA GB300 NVL72 GPU 288GB vs NVIDIA H100 SXM5 80GB

The GB300 delivers 2.5x the FP8 throughput of the H100 (5,000 vs 1,979 TFLOPS dense). The GB300 also carries 208 GB more memory (288 GB vs 80 GB).

GB300: Blackwell Ultra, 2025 H100: Hopper, 2022
FP8 dense lead
2.5x
GB300: 5,000 vs 1,979 TFLOPS
Memory
288 vs 80 GB
208 GB more for the GB300
Bandwidth
8.0 TB/s vs 3.4 TB/s
GB300 moves data faster
See rental pricing from $1.73/hr Download comparison sheet

Share this page

NVIDIA GB300 NVL72 GPU 288GB vs NVIDIA H100 SXM5 80GB

A post with this link shows the card on the Image tab as its preview.

Post on X Share on LinkedIn
 Printable version →

NVIDIA GB300 NVL72 GPU 288GB

Leads on FP8 Full specs →

What is the NVIDIA GB300?

The GB300 is a GPU from NVIDIA on the Blackwell Ultra architecture, an accelerator for AI training, inference and high-performance computing. It is an SXM module, the mezzanine format used in dense multi-GPU servers. GB300 full specifications →

Can I rent the GB300 in the cloud?

Yes. Cloud providers tracked by Flopper rent the GB300 by the hour, and the pricing page compares their current rates and the price history side by side. GB300 rental prices →

How does the GB300 compare with other accelerators?

This page sets the GB300 against the H100. The GB300 is the newer generation of the two. Both are NVIDIA parts, and both ship as an SXM module, so they fit the same class of system. Flopper also compares it head to head with the GB200 NVL72 GPU 186GB, across every published precision, memory and power. GB300 vs GB200 NVL72 GPU 186GB →

Where do the GB300 figures come from?

Every GB300 figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

NVIDIA H100 SXM5 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 an SXM module, the mezzanine format used in dense multi-GPU servers. 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 GB300. The H100 is the older generation of the two. Both are NVIDIA parts, and both ship as an SXM module, so they fit the same class of system. Flopper also compares it head to head with the A100 SXM4 80GB, H200 SXM 141GB and Instinct MI300X 192GB, across every published precision, memory and power. H100 vs A100 SXM4 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

GB300
H100
Architecture
Blackwell UltraHopper
Launch Year
20252022
Form Factor
SXMSXM
Memory
288 GB80 GB
Memory Bandwidth
8.0 TB/s3.4 TB/s
TDP
—700 W
Process Node
4nm4nm

Performance (TFLOPS)

GB300
H100
1.4 TFLOPS 34 TFLOPS
83 TFLOPS 67 TFLOPS
1,250 TFLOPS
2,500 TFLOPS sparse
494 TFLOPS
989 TFLOPS sparse
2,500 TFLOPS
5,000 TFLOPS sparse
990 TFLOPS
1,979 TFLOPS sparse
2,500 TFLOPS
5,000 TFLOPS sparse
990 TFLOPS
1,979 TFLOPS sparse
5,000 TFLOPS
10,000 TFLOPS sparse
1,979 TFLOPS
3,958 TFLOPS sparse
FP6
5,000 TFLOPS
10,000 TFLOPS sparse
 No verified data available
15,000 TFLOPS
20,000 TFLOPS sparse
 No verified data available
165 TOPS
330 TOPS sparse
1,979 TOPS
3,958 TOPS sparse
Tensor Core
No verified data available67 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 GB300 is the newer part: Blackwell Ultra (2025) against the H100's Hopper (2022).

The GB300 carries 208 GB more memory (288 GB against 80 GB), so larger models fit on one GPU before you shard. Bandwidth is 8.0 TB/s against 3.4 TB/s, which sets the ceiling on decode speed.

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, FP4, INT8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

What is better, the GB300 or the H100?

The GB300, on throughput. At FP8 precision the GB300 reaches 5,000 TFLOPS dense against 1,979 TFLOPS for the H100. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the GB300 or the H100?

Choose the GB300 for training and high-throughput inference: it has 2.5x the FP8 throughput of the H100. It also carries 208 GB more memory, so larger models fit on one GPU before you shard.

Which has more memory, the GB300 or the H100?

The GB300 carries 288 GB of memory versus 80 GB for the H100. Memory bandwidth is 8.0 TB/s for the GB300 and 3.4 TB/s for the H100.

How much do the GB300 and the H100 cost to rent?

Live cloud listings tracked by Flopper price the GB300 from $5.74 per GPU-hour across 2 providers, and the H100 from $1.73 per GPU-hour across 34 providers. The rental section on this page links to every provider, rate and the price history for each GPU.

Where to Rent

NVIDIA GB300 NVL72 GPU 288GB

Rent from

$5.74 / GPU-hr

Lowest of 3 live listings across 2 providers

See all GB300 rental prices →

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

NVIDIA H100 SXM5 80GB

Rent from

$1.73 / GPU-hr

Lowest of 66 live listings across 34 providers · median $3.85

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 → GB300 full specs H100 full specs Estimate cluster cost in the calculator →

© 2026 Flopper.io - Compare the GPUs Powering AI