readygpu

NVIDIA GeForce RTX 3090 vs NVIDIA RTX PRO 2000 Blackwell

For running models locally, priced in the US. At 8k context the first runs 61 models fully in memory, the second 44.

In short

The NVIDIA GeForce RTX 3090 has more memory for models: 24 GB against 16 GB. At 8k context the NVIDIA GeForce RTX 3090 runs 61 of the models we track fully in memory and the NVIDIA RTX PRO 2000 Blackwell 44; only the NVIDIA GeForce RTX 3090 holds Qwen 3.8 27B. On the models both run fully in memory, the NVIDIA GeForce RTX 3090 is about 3.2× as fast as the NVIDIA RTX PRO 2000 Blackwell (estimated, median across models). In the US the NVIDIA GeForce RTX 3090 costs $1,698 at Amazon and the NVIDIA RTX PRO 2000 Blackwell $1,580 at eBay: the NVIDIA RTX PRO 2000 Blackwell is $118 cheaper.

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

NVIDIA GeForce RTX 3090

$1,698

Amazon

Add to build
NVIDIA RTX PRO 2000 Blackwell

$1,580

eBay

Add to build

Specifications

SpecNVIDIA GeForce RTX 3090NVIDIA RTX PRO 2000 Blackwell
GPUGeForce RTX 3090RTX PRO 2000 Blackwell
Memory24 GB GDDR6X16 GB GDDR7
Memory bus384-bit128-bit
Memory bandwidth936 GB/s288 GB/s
FP16 compute71 TFLOPS17.03 TFLOPS
SoftwareCUDACUDA
Board power350 W70 W
PCIe generationGen 4Gen 5
PCIe lanesx16x8

What each runs

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

ModelNVIDIA GeForce RTX 3090NVIDIA RTX PRO 2000 Blackwell
Qwen 3.8 27B26tok/sUD-Q6_Koffload
Qwen 3.6 27B25tok/sQ6_Koffload
Gemma 4 31B25tok/sQ5_K_Moffload
Qwen 3.6 35B-A3B167tok/sQ4_K_Moffload
Qwen 3.5 27B25tok/sQ6_Koffload
Gemma 4 26B-A4B88tok/sUD-Q6_Koffload
Qwen 3.5 35B-A3B168tok/sQ4_K_Moffload
GLM 4.7 Flash87tok/sQ5_K_Moffload
Devstral Small 2 24B28tok/sQ6_K11tok/sQ4_K_M
Nemotron 3.5 Lightning 30B-A3B151tok/sMXFP4offload
gpt-oss 20B132tok/sMXFP441tok/sMXFP4
Qwen3 Coder 30B-A3B118tok/sQ5_K_Moffload
Qwen3 32B27tok/sQ4_K_Moffload
Gemma 4 12B44tok/sQ8_014tok/sQ8_0
Qwen 3.5 9B59tok/sQ8_018tok/sQ8_0
Mistral Small 3.2 24B28tok/sQ6_K11tok/sQ4_K_M
DeepSeek R1 Distill Qwen 32B26tok/sQ4_K_Moffload
Gemma 3 27B25tok/sQ6_Koffload
Ministral 3 14B37tok/sQ8_011tok/sQ8_0
Qwen3 14B34tok/sQ8_013tok/sQ6_K
DeepSeek R1 0528 Qwen3 8B59tok/sQ8_018tok/sQ8_0
Qwen3 8B59tok/sQ8_018tok/sQ8_0
Phi-4 14B34tok/sQ8_013tok/sQ6_K
Qwen 3.5 4B121tok/sQ8_037tok/sQ8_0
Ministral 3 8B57tok/sQ8_018tok/sQ8_0
Granite 4.2 8B54tok/sQ8_017tok/sQ8_0
Gemma 4 E4B69tok/sQ8_021tok/sQ8_0
LFM2.5 8B-A1B237tok/sQ8_073tok/sQ8_0
Llama 3.1 8B76tok/sQ6_K23tok/sQ6_K
Qwen2.5 0.5B919tok/sQ8_0283tok/sQ8_0
Qwen3.5 0.8B623tok/sQ8_0192tok/sQ8_0
Qwen3.5 2B373tok/sQ5_K_M115tok/sQ5_K_M
LFM3 1.7B297tok/sQ8_091tok/sQ8_0
DeepSeek R1 Distill Qwen 1.5B273tok/sQ8_084tok/sQ8_0
MiniCPM5 2B191tok/sQ8_059tok/sQ8_0
LFM2.5 2.6B172tok/sQ8_053tok/sQ8_0
LFM2.5 VL 3B172tok/sQ8_053tok/sQ8_0
Spark X2.5 4B171tok/smlx-4bit53tok/smlx-4bit
LFM2.5 24B A2B143tok/sQ6_Koffload
granite vision 4.1 4b135tok/sQ8_042tok/sQ8_0
Ministral 3 3B Instruct 2512128tok/sQ8_040tok/sQ8_0
granite 4.2 3b127tok/sQ8_039tok/sQ8_0
Mellum2.1 12B A2.5B Thinking123tok/sQ8_038tok/sQ8_0
FrogNano 4B 2609118tok/sQ8_036tok/sQ8_0
Kimi VL A3B Thinking 2506104tok/sQ8_049tok/sQ4_K_M
Kimi VL A3B99tok/smlx-8bit39tok/smlx-6bit
DeepSeek R1 Distill Qwen 7B68tok/sQ8_021tok/sQ8_0
GLM 4 9B 041468tok/smlx-5bit21tok/smlx-5bit
DeepSeek R1 Distill Llama 8B60tok/sQ8_019tok/sQ8_0
Ministral 3 8B Reasoning 251257tok/sQ8_018tok/sQ8_0
GLM 4.6V Flash56tok/sQ8_017tok/sQ8_0
GLM 4.1V 9B Thinking48tok/smlx-8bit15tok/smlx-8bit
Ministral 3 14B Reasoning 251237tok/sQ8_011tok/sQ8_0
DeepSeek R1 Distill Qwen 14B34tok/sQ8_013tok/sQ6_K
Devstral Small 250732tok/sQ5_K_M11tok/sQ4_K_M
Magistral Small 250732tok/sQ5_K_M11tok/sQ4_K_M
Magistral Small 250932tok/sQ5_K_M11tok/sQ4_K_M
Magistral Small 250628tok/sQ6_K11tok/sQ4_K_M
Qwen3.8 27B OBLITERATED27tok/smlx-4bitoffload
granite 4.2 30b25tok/sQ5_K_Moffload
Muse Glimmer 30B25tok/sQ6_Koffload
Qwen 3.5 122B-A10Boffloadoffload
Qwen3 Coder Nextoffloadwon't run
Mistral Small 4 119Boffloadoffload
GLM 4.5 Airoffloadwon't run
Qwen3 Next 80B-A3Boffloadwon't run
Kimi Linear 48B-A3Boffloadoffload
Llama 3.3 70Boffloadwon't run
DeepSeek R1 Distill Llama 70Boffloadwon't run
DeepSeek V4 Flashwon't runwon't run
GLM 5.3 Flashwon't runwon't run
Qwen 3.8 Flash Nextwon't runwon't run
MiniMax M2.7won't runwon't run
Nemotron 3 Super 120B-A12Bwon't runwon't run
gpt-oss 120Bwon't runwon't run
DeepSeek R1won't runwon't run
DeepSeek R1 0528won't runwon't run
DeepSeek R1 Zerowon't runwon't run
DeepSeek V3won't runwon't run
DeepSeek V3 0324won't runwon't run
DeepSeek V3.1won't runwon't run
DeepSeek V3.1 Terminuswon't runwon't run
DeepSeek V3.2won't runwon't run
DeepSeek V4 Flash 0731won't runwon't run
DeepSeek V4 Flash Vision Expwon't runwon't run
Devstral 2 123B Instruct 2512won't runwon't run
GLM 4.5won't runwon't run
GLM 4.5Vwon't runwon't run
GLM 4.6won't runwon't run
GLM 4.6Vwon't runwon't run
GLM 4.7won't runwon't run
GLM 5won't runwon't run
GLM 5.1won't runwon't run
GLM 5.2won't runwon't run
GLM 5.3won't runwon't run
Kimi Dev 72Bwon't runwon't run
Kimi K2won't runwon't run
Kimi K2 Instruct 0905won't runwon't run
Kimi K2 Thinkingwon't runwon't run
Kimi K2.5won't runwon't run
MiniMax M2won't runwon't run
MiniMax M2.1won't runwon't run
MiniMax M2.5won't runwon't run
MiniMax M3won't runwon't run
Mistral Medium 3.5 128Bwon't runwon't run

Questions

Which is better for local AI, the NVIDIA GeForce RTX 3090 or the NVIDIA RTX PRO 2000 Blackwell?
Choose the NVIDIA GeForce RTX 3090 to run bigger models. Choose the NVIDIA GeForce RTX 3090 for speed. Choose the NVIDIA RTX PRO 2000 Blackwell to spend less.
Which can run bigger models, the NVIDIA GeForce RTX 3090 or the NVIDIA RTX PRO 2000 Blackwell?
At 8k context the NVIDIA GeForce RTX 3090 runs 61 of the models we track fully in memory and the NVIDIA RTX PRO 2000 Blackwell 44; only the NVIDIA GeForce RTX 3090 holds Qwen 3.8 27B.
Which is faster for local AI, the NVIDIA GeForce RTX 3090 or the NVIDIA RTX PRO 2000 Blackwell?
On the models both run fully in memory, the NVIDIA GeForce RTX 3090 is about 3.2× as fast as the NVIDIA RTX PRO 2000 Blackwell (estimated, median across models).
What is the price difference between the NVIDIA GeForce RTX 3090 and the NVIDIA RTX PRO 2000 Blackwell?
In the US the NVIDIA GeForce RTX 3090 costs $1,698 at Amazon and the NVIDIA RTX PRO 2000 Blackwell $1,580 at eBay: the NVIDIA RTX PRO 2000 Blackwell is $118 cheaper.

More comparisons

More hardware for local AI