readygpu

NVIDIA RTX PRO 2000 Blackwell vs NVIDIA Tesla P40

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

In short

The NVIDIA Tesla P40 has more memory for models: 24 GB against 16 GB. At 8k context the NVIDIA Tesla P40 runs 61 of the models we track fully in memory and the NVIDIA RTX PRO 2000 Blackwell 44; only the NVIDIA Tesla P40 holds Qwen 3.8 27B. On the models both run fully in memory, the NVIDIA Tesla P40 is about 1.2× as fast as the NVIDIA RTX PRO 2000 Blackwell (estimated, median across models).

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

NVIDIA RTX PRO 2000 Blackwell

$2,241

Amazon

Add to build
NVIDIA Tesla P40

Not priced here

Add to build

Specifications

SpecNVIDIA RTX PRO 2000 BlackwellNVIDIA Tesla P40
GPURTX PRO 2000 BlackwellTesla P40
Memory16 GB GDDR724 GB GDDR5
Memory bus128-bit384-bit
Memory bandwidth288 GB/s346 GB/s
FP16 compute17.03 TFLOPS11.8 TFLOPS
SoftwareCUDACUDA
Board power70 W250 W
PCIe generationGen 5Gen 3
PCIe lanesx8x16

What each runs

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

ModelNVIDIA RTX PRO 2000 BlackwellNVIDIA Tesla P40
Qwen 3.8 27Boffload10tok/sUD-Q6_K
Qwen 3.6 27Boffload9tok/sQ6_K
Gemma 4 31Boffload9tok/sQ5_K_M
Qwen 3.6 35B-A3Boffload62tok/sQ4_K_M
Qwen 3.5 27Boffload9tok/sQ6_K
Gemma 4 26B-A4Boffload32tok/sUD-Q6_K
Devstral Small 2 24B11tok/sQ4_K_M10tok/sQ6_K
Qwen 3.5 35B-A3Boffload62tok/sQ4_K_M
GLM 4.7 Flashoffload32tok/sQ5_K_M
Nemotron 3.5 Lightning 30B-A3Boffload56tok/sMXFP4
gpt-oss 20B41tok/sMXFP449tok/sMXFP4
Qwen3 Coder 30B-A3Boffload44tok/sQ5_K_M
Qwen3 32Boffload10tok/sQ4_K_M
Gemma 4 12B14tok/sQ8_016tok/sQ8_0
Qwen 3.5 9B18tok/sQ8_022tok/sQ8_0
Mistral Small 3.2 24B11tok/sQ4_K_M10tok/sQ6_K
Gemma 3 27Boffload9tok/sQ6_K
DeepSeek R1 Distill Qwen 32Boffload10tok/sQ4_K_M
Qwen3 14B13tok/sQ6_K13tok/sQ8_0
Ministral 3 14B11tok/sQ8_014tok/sQ8_0
DeepSeek R1 0528 Qwen3 8B18tok/sQ8_022tok/sQ8_0
Qwen3 8B18tok/sQ8_022tok/sQ8_0
Phi-4 14B13tok/sQ6_K12tok/sQ8_0
Qwen 3.5 4B37tok/sQ8_045tok/sQ8_0
Ministral 3 8B18tok/sQ8_021tok/sQ8_0
Granite 4.2 8B17tok/sQ8_020tok/sQ8_0
Gemma 4 E4B21tok/sQ8_025tok/sQ8_0
LFM2.5 8B-A1B73tok/sQ8_088tok/sQ8_0
Llama 3.1 8B23tok/sQ6_K28tok/sQ6_K
Qwen2.5 0.5B283tok/sQ8_0340tok/sQ8_0
Qwen3.5 0.8B192tok/sQ8_0230tok/sQ8_0
Qwen3.5 2B115tok/sQ5_K_M138tok/sQ5_K_M
LFM3 1.7B91tok/sQ8_0110tok/sQ8_0
DeepSeek R1 Distill Qwen 1.5B84tok/sQ8_0101tok/sQ8_0
MiniCPM5 2B59tok/sQ8_071tok/sQ8_0
LFM2.5 2.6B53tok/sQ8_064tok/sQ8_0
LFM2.5 VL 3B53tok/sQ8_064tok/sQ8_0
Spark X2.5 4B53tok/smlx-4bit63tok/smlx-4bit
Kimi VL A3B Thinking 250649tok/sQ4_K_M38tok/sQ8_0
granite vision 4.1 4b42tok/sQ8_050tok/sQ8_0
Ministral 3 3B Instruct 251240tok/sQ8_047tok/sQ8_0
granite 4.2 3b39tok/sQ8_047tok/sQ8_0
Kimi VL A3B39tok/smlx-6bit37tok/smlx-8bit
Mellum2.1 12B A2.5B Thinking38tok/sQ8_046tok/sQ8_0
FrogNano 4B 260936tok/sQ8_043tok/sQ8_0
DeepSeek R1 Distill Qwen 7B21tok/sQ8_025tok/sQ8_0
GLM 4 9B 041421tok/smlx-5bit25tok/smlx-5bit
DeepSeek R1 Distill Llama 8B19tok/sQ8_022tok/sQ8_0
Ministral 3 8B Reasoning 251218tok/sQ8_021tok/sQ8_0
GLM 4.6V Flash17tok/sQ8_021tok/sQ8_0
GLM 4.1V 9B Thinking15tok/smlx-8bit18tok/smlx-8bit
DeepSeek R1 Distill Qwen 14B13tok/sQ6_K12tok/sQ8_0
Devstral Small 250711tok/sQ4_K_M12tok/sQ5_K_M
Magistral Small 250611tok/sQ4_K_M10tok/sQ6_K
Magistral Small 250711tok/sQ4_K_M12tok/sQ5_K_M
Magistral Small 250911tok/sQ4_K_M12tok/sQ5_K_M
Ministral 3 14B Reasoning 251211tok/sQ8_014tok/sQ8_0
LFM2.5 24B A2Boffload53tok/sQ6_K
Muse Glimmer 30Boffload9tok/sQ6_K
granite 4.2 30boffload9tok/sQ5_K_M
Qwen3.8 27B OBLITERATEDoffload10tok/smlx-4bit
Qwen 3.5 122B-A10Boffloadoffload
Qwen3 Coder Nextwon't runoffload
Mistral Small 4 119Boffloadoffload
GLM 4.5 Airwon't runoffload
Qwen3 Next 80B-A3Bwon't runoffload
Kimi Linear 48B-A3Boffloadoffload
Llama 3.3 70Bwon't runoffload
DeepSeek R1 Distill Llama 70Bwon't runoffload
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 RTX PRO 2000 Blackwell or the NVIDIA Tesla P40?
Choose the NVIDIA Tesla P40 to run bigger models. Choose the NVIDIA Tesla P40 for speed.
Which can run bigger models, the NVIDIA RTX PRO 2000 Blackwell or the NVIDIA Tesla P40?
At 8k context the NVIDIA Tesla P40 runs 61 of the models we track fully in memory and the NVIDIA RTX PRO 2000 Blackwell 44; only the NVIDIA Tesla P40 holds Qwen 3.8 27B.
Which is faster for local AI, the NVIDIA RTX PRO 2000 Blackwell or the NVIDIA Tesla P40?
On the models both run fully in memory, the NVIDIA Tesla P40 is about 1.2× as fast as the NVIDIA RTX PRO 2000 Blackwell (estimated, median across models).

More comparisons

More hardware for local AI