readygpu

NVIDIA GeForce RTX 5090 vs NVIDIA RTX PRO 6000 Blackwell

For running models locally, priced in India. At 64k context the first runs 27 models fully in memory, the second 40.

In short

The NVIDIA RTX PRO 6000 Blackwell has more memory for models: 96 GB against 32 GB. At 64k context the NVIDIA RTX PRO 6000 Blackwell runs 40 of the models we track fully in memory and the NVIDIA GeForce RTX 5090 27; only the NVIDIA RTX PRO 6000 Blackwell holds Qwen 3.8 Flash Next. On the models both run fully in memory, the NVIDIA GeForce RTX 5090 and the NVIDIA RTX PRO 6000 Blackwell are about as fast as each other (estimated).

Updated . Speeds are estimates; how they are worked out.

NVIDIA GeForce RTX 5090

Not priced here

Add to build
NVIDIA RTX PRO 6000 Blackwell

Not priced here

Add to build

Specifications

SpecNVIDIA GeForce RTX 5090NVIDIA RTX PRO 6000 Blackwell
GPUGeForce RTX 5090RTX PRO 6000 Blackwell
Memory32 GB GDDR796 GB GDDR7
Memory bus512-bit512-bit
Memory bandwidth1,792 GB/s1,792 GB/s
FP16 compute209 TFLOPS504 TFLOPS
SoftwareCUDACUDA
Board power575 W600 W
PCIe generationGen 5Gen 5
PCIe lanesx16x16

What each runs

Estimated decode speed with the most precise quant that fits each one est.

ModelNVIDIA GeForce RTX 5090NVIDIA RTX PRO 6000 Blackwell
Qwen 3.8 Flash Nextwon't run21tok/sUD-Q3_K_XL
Qwen 3.8 27B42tok/sUD-Q6_K33tok/sQ8_0
MiniMax M2.7won't run33tok/sUD-Q2_K_XL
Qwen 3.5 122B-A10Boffload73tok/sQ5_K_M
Qwen 3.6 27B41tok/sQ6_K34tok/sQ8_0
Gemma 4 31B35tok/sQ6_K29tok/sQ8_0
Qwen3 Coder Nextoffload115tok/sQ8_0
Nemotron 3 Super 120B-A12Boffload77tok/sUD-Q4_K_M
gpt-oss 120Bwon't run120tok/sMXFP4
Mistral Small 4 119Boffload90tok/sUD-Q6_K
Qwen 3.6 35B-A3B173tok/sUD-Q6_K149tok/sQ8_0
Qwen 3.5 27B41tok/sQ6_K34tok/sQ8_0
Gemma 4 26B-A4B117tok/sQ8_0117tok/sQ8_0
GLM 4.5 Airoffload26tok/sQ5_K_M
Qwen 3.5 35B-A3B175tok/sQ6_K149tok/sQ8_0
GLM 4.7 Flash87tok/sQ6_K76tok/sQ8_0
Devstral Small 2 24B37tok/sQ6_K31tok/sQ8_0
Qwen3 Next 80B-A3Boffload139tok/sQ8_0
Nemotron 3.5 Lightning 30B-A3Boffload177tok/sUD-Q6_K_XL
Kimi Linear 48B-A3B277tok/sQ4_K_M177tok/sQ8_0
gpt-oss 20B165tok/sMXFP4165tok/sMXFP4
Qwen3 Coder 30B-A3B75tok/sQ6_K69tok/sQ8_0
Qwen3 32Boffload21tok/sQ8_0
Gemma 4 12B79tok/sQ8_079tok/sQ8_0
Qwen 3.5 9B95tok/sQ8_095tok/sQ8_0
Gemma 3 27B40tok/sQ6_K32tok/sQ8_0
Mistral Small 3.2 24B37tok/sQ6_K31tok/sQ8_0
DeepSeek R1 Distill Qwen 32Boffload21tok/sQ8_0
Llama 3.3 70Bwon't run12tok/sQ8_0
Ministral 3 14B44tok/sQ8_044tok/sQ8_0
Qwen3 14B42tok/sQ8_042tok/sQ8_0
DeepSeek R1 0528 Qwen3 8B61tok/sQ8_061tok/sQ8_0
Qwen3 8B61tok/sQ8_061tok/sQ8_0
Phi-4 14B38tok/sQ8_038tok/sQ8_0
Qwen 3.5 4B166tok/sQ8_0166tok/sQ8_0
Ministral 3 8B61tok/sQ8_061tok/sQ8_0
Granite 4.2 8B55tok/sQ8_055tok/sQ8_0
Gemma 4 E4B110tok/sQ8_0110tok/sQ8_0
LFM2.5 8B-A1B156tok/sQ8_0156tok/sQ8_0
Llama 3.1 8B73tok/sQ6_K73tok/sQ6_K
DeepSeek V4 Flashwon't runoffload
GLM 5.3 Flashwon't runoffload

Questions

Which is better for local AI, the NVIDIA GeForce RTX 5090 or the NVIDIA RTX PRO 6000 Blackwell?
Choose the NVIDIA RTX PRO 6000 Blackwell to run bigger models.
Which can run bigger models, the NVIDIA GeForce RTX 5090 or the NVIDIA RTX PRO 6000 Blackwell?
At 64k context the NVIDIA RTX PRO 6000 Blackwell runs 40 of the models we track fully in memory and the NVIDIA GeForce RTX 5090 27; only the NVIDIA RTX PRO 6000 Blackwell holds Qwen 3.8 Flash Next.
Which is faster for local AI, the NVIDIA GeForce RTX 5090 or the NVIDIA RTX PRO 6000 Blackwell?
On the models both run fully in memory, the NVIDIA GeForce RTX 5090 and the NVIDIA RTX PRO 6000 Blackwell are about as fast as each other (estimated).

More comparisons