NVIDIA GB10 Grace Blackwell vs NVIDIA RTX PRO 6000 Blackwell Workstation Edition

Both GPUs launched in 2025: the GB10 Grace Blackwell on Blackwell and the RTX PRO 6000 Blackwell on Blackwell.

Dense throughput for the GB10 Grace Blackwell against the RTX PRO 6000 Blackwell: FP32 31 vs 126 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 128 GB against 96 GB, fed at 273 GB/s versus 1.8 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 180 W for the GB10 Grace Blackwell and 600 W for the RTX PRO 6000 Blackwell. At FP32 that works out to 0.17 against 0.21 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.