Flopper.io Flopper.io

NVIDIA B100 SXM 192GB vs NVIDIA GB200 Grace Blackwell Superchip 372GB

The GB200 delivers 2.9x the FP8 throughput of the B100 (10,000 vs 3,500 TFLOPS dense). The GB200 also carries 180 GB more memory (372 GB vs 192 GB).

B100: Blackwell, 2024 GB200: Blackwell, 2024
FP8 dense lead
2.9x
GB200: 10,000 vs 3,500 TFLOPS
Memory
192 vs 372 GB
180 GB more for the GB200
Bandwidth
8.0 TB/s vs 16.0 TB/s
GB200 moves data faster
Download comparison sheet

Share this page

NVIDIA B100 SXM 192GB vs NVIDIA GB200 Grace Blackwell Superchip 372GB

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

Post on X Share on LinkedIn
 Printable version →

NVIDIA B100 SXM 192GB

Full specs →

What is the NVIDIA B100?

The B100 is a GPU from NVIDIA on the Blackwell 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. B100 full specifications →

Can I rent the B100 in the cloud?

Flopper is not tracking any live rental listings for the B100 yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the B100 compare with other accelerators?

This page sets the B100 against the GB200. Both parts launched in the same year. Both are NVIDIA parts; the B100 is an SXM module while the GB200 is a superchip. Flopper also compares it head to head with the B200 SXM 180GB, across every published precision, memory and power. B100 vs B200 SXM 180GB →

Where do the B100 figures come from?

NVIDIA published the B100 figures in marketing material rather than a datasheet, and Flopper has no independent confirmation of them. How Flopper verifies specifications →

NVIDIA GB200 Grace Blackwell Superchip 372GB

Leads on FP8 Full specs →

What is the NVIDIA GB200?

The GB200 is a GPU from NVIDIA on the Blackwell 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. GB200 full specifications →

Can I rent the GB200 in the cloud?

Flopper is not tracking any live rental listings for the GB200 yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the GB200 compare with other accelerators?

This page sets the GB200 against the B100. Both parts launched in the same year. Both are NVIDIA parts; the GB200 is a superchip while the B100 is an SXM module. The compare tool builds the same comparison for any two parts in the catalogue. All GPU comparisons →

Where do the GB200 figures come from?

Every GB200 figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

Specifications

B100
GB200
Architecture
BlackwellBlackwell
Launch Year
20242024
Form Factor
SXMSuperchip
Memory
192 GB372 GB
Memory Bandwidth
8.0 TB/s16.0 TB/s
TDP
700 W—
Process Node
4nm4nm

Performance (TFLOPS)

B100
GB200
30 TFLOPS 80 TFLOPS
60 TFLOPS 160 TFLOPS
875 TFLOPS
sparse not published
2,500 TFLOPS
5,000 TFLOPS sparse
1,750 TFLOPS
sparse not published
5,000 TFLOPS
10,000 TFLOPS sparse
1,750 TFLOPS
sparse not published
5,000 TFLOPS
10,000 TFLOPS sparse
3,500 TFLOPS
sparse not published
10,000 TFLOPS
20,000 TFLOPS sparse
FP6
No verified data available10,000 TFLOPS
20,000 TFLOPS sparse
7,000 TFLOPS
sparse not published
20,000 TFLOPS
40,000 TFLOPS sparse
3,500 TOPS
sparse not published
10,000 TOPS
20,000 TOPS sparse

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

Both launched in 2024, the B100 on Blackwell and the GB200 on Blackwell.

The GB200 carries 180 GB more memory (372 GB against 192 GB), so larger models fit on one GPU before you shard. Bandwidth is 8.0 TB/s against 16.0 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

Is the B100 faster than the GB200?

At FP8 precision the GB200 reaches 10,000 TFLOPS dense against 3,500 TFLOPS for the B100. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the B100 or the GB200?

Choose the GB200 for training and high-throughput inference: it has 2.9x the FP8 throughput of the B100. It also carries 180 GB more memory, so larger models fit on one GPU before you shard. Packaging differs too: the B100 is an SXM module while the GB200 is a superchip, so check what your servers accept.

Which has more memory, the B100 or the GB200?

The GB200 carries 372 GB of memory versus 192 GB for the B100. Memory bandwidth is 8.0 TB/s for the B100 and 16.0 TB/s for the GB200.

Get Comparison Updates

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

Open in Full Compare Tool → B100 full specs GB200 full specs Estimate cluster cost in the calculator →

© 2026 Flopper.io - Compare the GPUs Powering AI