NVIDIA B200 SXM 180GB vs NVIDIA GB200 Grace Blackwell Superchip 372GB

Both GPUs launched in 2024: the B200 on Blackwell and the GB200 on Blackwell.

Dense throughput for the B200 against the GB200: FP64 37 vs 80 TFLOPS, FP32 75 vs 160 TFLOPS, FP16 2,250 vs 5,000 TFLOPS, FP8 4,500 vs 10,000 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 180 GB against 372 GB, fed at 7.7 TB/s versus 16.0 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.

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.