NVIDIA RTX PRO 6000 Blackwell Max-Q Workstation Edition vs NVIDIA GeForce RTX 5090 32GB

Both GPUs launched in 2025: the RTX PRO 6000 Blackwell on Blackwell and the GeForce RTX 5090 on Blackwell.

Dense throughput for the RTX PRO 6000 Blackwell against the GeForce RTX 5090: FP64 1.7 vs 1.6 TFLOPS, FP32 110 vs 105 TFLOPS, FP16 439 vs 419 TFLOPS, FP8 878 vs 838 TFLOPS. Sparse figures, where the vendor publishes them, appear under each dense number in the performance table.

Memory is 96 GB against 32 GB, fed at 1.8 TB/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 300 W for the RTX PRO 6000 Blackwell and 575 W for the GeForce RTX 5090. At FP32 that works out to 0.37 against 0.18 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.