NVIDIA B100 SXM 192GB vs NVIDIA GB10 Grace Blackwell

The B100 delivers 94% more FP32 throughput than the GB10 Grace Blackwell (60 vs 31 TFLOPS dense). The B100 also carries 64 GB more memory (192 GB vs 128 GB).

B100: Blackwell, 2024 GB10 Grace Blackwell: Blackwell, 2025
FP32 dense lead
+94%
B100: 60 vs 31 TFLOPS
Memory
192 vs 128 GB
64 GB more for the B100
Bandwidth
8.0 TB/s vs 273 GB/s
B100 moves data faster
TDP
700 W vs 180 W
GB10 Grace Blackwell draws 520 W less
Download comparison sheet

Share this page

NVIDIA B100 SXM 192GB vs NVIDIA GB10 Grace Blackwell

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

Printable version →

Specifications

B100
GB10 Grace Blackwell
Architecture
BlackwellBlackwell
Launch Year
20242025
Form Factor
SXMIntegrated
Memory
192 GB128 GB
Memory Bandwidth
8.0 TB/s273 GB/s
TDP
700 W180 W
Process Node
4nm—

Performance (TFLOPS)

B100
GB10 Grace Blackwell
FP64
30 TFLOPS No verified data available
FP32
60 TFLOPS 31 TFLOPS
TF32
875 TFLOPS
sparse not published
No verified data available
BF16
1,750 TFLOPS
sparse not published
No verified data available
FP16
1,750 TFLOPS
sparse not published
No verified data available
FP8
3,500 TFLOPS
sparse not published
No verified data available
FP6
No verified data available No verified data available
FP4
7,000 TFLOPS
sparse not published
500 TFLOPS
1,000 TFLOPS sparse
INT8
3,500 TOPS
sparse not published
No verified data available

FLOPS by Precision

What actually differs

The GB10 Grace Blackwell is the newer part: Blackwell, launched in 2025, against the B100's Blackwell from 2024. Newer architectures typically add lower-precision formats and better throughput per watt, so check the precision rows your workload actually uses.

Dense throughput for the B100 against the GB10 Grace Blackwell: FP32 60 vs 31 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 192 GB against 128 GB, fed at 8.0 TB/s versus 273 GB/s. More memory per GPU means larger models fit before you have to shard across cards, which often matters more than raw TFLOPS for inference.

Power budgets are 700 W for the B100 and 180 W for the GB10 Grace Blackwell. At FP32 that works out to 0.09 against 0.17 TFLOPS per watt.

For LLM training and inference, weight the FP8 and FP16 rows and memory capacity most heavily. For scientific and HPC workloads, the FP64 row is the one to read.

Frequently asked questions

Is the B100 faster than the GB10 Grace Blackwell?

At FP32 precision the B100 reaches 60 TFLOPS dense against 31 TFLOPS for the GB10 Grace Blackwell. The performance table on this page lists every published precision for both GPUs.

Which has more memory, the B100 or the GB10 Grace Blackwell?

The B100 carries 192 GB of memory versus 128 GB for the GB10 Grace Blackwell. Memory bandwidth is 8.0 TB/s for the B100 and 273 GB/s for the GB10 Grace Blackwell.

How much power do the B100 and the GB10 Grace Blackwell draw?

The B100 is rated at 700 W TDP and the GB10 Grace Blackwell at 180 W. On FP32 throughput per watt, the GB10 Grace Blackwell is the more efficient part.

Get Comparison Updates

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

© 2026 Flopper.io - Compare the GPUs Powering AI