NVIDIA B200 SXM 180GB vs NVIDIA B100 SXM 192GB

Both GPUs launched in 2024: the B200 on Blackwell and the B100 on Blackwell.

Dense throughput for the B200 against the B100: FP64 37 vs 30 TFLOPS, FP32 75 vs 60 TFLOPS, FP16 2,250 vs 1,750 TFLOPS, FP8 4,500 vs 3,500 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 180 GB against 192 GB, fed at 7.7 TB/s versus 8.0 TB/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 1.0 kW for the B200 and 700 W for the B100. At FP32 that works out to 0.07 against 0.09 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.