readygpu

AMD Radeon AI PRO R9700 vs NVIDIA RTX PRO 4500 Blackwell

For running models locally, priced in the US. At 128k context the first runs 50 models fully in memory, the second 50.

In short

Both have 32 GB of memory for models; the AMD Radeon AI PRO R9700 has 640 GB/s of bandwidth and the NVIDIA RTX PRO 4500 Blackwell 896 GB/s. At 128k context both run the same 50 models we track fully in memory. On the models both run fully in memory, the NVIDIA RTX PRO 4500 Blackwell is about 1.6× as fast as the AMD Radeon AI PRO R9700 (estimated, median across models).

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

AMD Radeon AI PRO R9700

$1,944

Amazon

Add to build
NVIDIA RTX PRO 4500 Blackwell

Not priced here

Add to build

Specifications

SpecAMD Radeon AI PRO R9700NVIDIA RTX PRO 4500 Blackwell
GPURadeon AI PRO R9700RTX PRO 4500 Blackwell
Memory32 GB GDDR632 GB GDDR7
Memory bus256-bit256-bit
Memory bandwidth640 GB/s896 GB/s
FP16 compute191 TFLOPS54.94 TFLOPS
SoftwareROCmCUDA
Board power300 W200 W
PCIe generationGen 5Gen 5
PCIe lanesx16x16

What each runs

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

ModelAMD Radeon AI PRO R9700NVIDIA RTX PRO 4500 Blackwell
Qwen 3.8 27B12tok/sUD-Q6_K18tok/sUD-Q6_K
Qwen 3.6 27B11tok/sQ6_K18tok/sQ6_K
Gemma 4 31B12tok/sQ4_K_M19tok/sQ4_K_M
Qwen 3.6 35B-A3B43tok/sUD-Q5_K_M64tok/sUD-Q6_K
Qwen 3.5 27B11tok/sQ6_K18tok/sQ6_K
Gemma 4 26B-A4B30tok/sQ8_047tok/sQ8_0
Qwen 3.5 35B-A3B41tok/sQ6_K65tok/sQ6_K
GLM 4.7 Flash19tok/sQ6_K30tok/sQ6_K
Nemotron 3.5 Lightning 30B-A3B57tok/sUD-Q5_K_M90tok/sUD-Q5_K_M
Kimi Linear 48B-A3B72tok/sQ4_K_M114tok/sQ4_K_M
gpt-oss 20B37tok/sMXFP459tok/sMXFP4
Qwen3 Coder 30B-A3B14tok/sQ4_K_M23tok/sQ4_K_M
Gemma 4 12B23tok/sQ8_037tok/sQ8_0
Qwen 3.5 9B25tok/sQ8_040tok/sQ8_0
Gemma 3 27B12tok/sQ5_K_M18tok/sQ5_K_M
Qwen3 14B12tok/sQ4_K_M17tok/sQ5_K_M
Ministral 3 14B11tok/sQ5_K_M18tok/sQ5_K_M
DeepSeek R1 0528 Qwen3 8B13tok/sQ8_020tok/sQ8_0
Qwen3 8B13tok/sQ8_020tok/sQ8_0
Qwen 3.5 4B40tok/sQ8_063tok/sQ8_0
Ministral 3 8B13tok/sQ8_020tok/sQ8_0
Granite 4.2 8B11tok/sQ8_018tok/sQ8_0
Gemma 4 E4B29tok/sQ8_046tok/sQ8_0
LFM2.5 8B-A1B28tok/sQ8_044tok/sQ8_0
Llama 3.1 8B15tok/sQ6_K23tok/sQ6_K
Qwen2.5 0.5B164tok/sQ8_0259tok/sQ8_0
Qwen3.5 0.8B144tok/sQ8_0228tok/sQ8_0
Qwen3.5 2B115tok/sQ5_K_M181tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B62tok/sQ8_098tok/sQ8_0
Mellum2.1 12B A2.5B Thinking46tok/sQ8_073tok/sQ8_0
Spark X2.5 4B44tok/smlx-4bit70tok/smlx-4bit
MiniCPM5 2B42tok/sQ8_067tok/sQ8_0
FrogNano 4B 260939tok/sQ8_062tok/sQ8_0
LFM2.5 2.6B32tok/sQ8_051tok/sQ8_0
LFM2.5 VL 3B32tok/sQ8_051tok/sQ8_0
Kimi VL A3B Thinking 250629tok/sQ8_046tok/sQ8_0
Kimi VL A3B29tok/smlx-8bit45tok/smlx-8bit
GLM 4 9B 041426tok/smlx-5bit41tok/smlx-5bit
granite vision 4.1 4b25tok/sQ8_039tok/sQ8_0
granite 4.2 3b24tok/sQ8_038tok/sQ8_0
GLM 4.6V Flash23tok/sQ8_036tok/sQ8_0
DeepSeek R1 Distill Qwen 7B23tok/sQ8_036tok/sQ8_0
GLM 4.1V 9B Thinking21tok/smlx-8bit32tok/smlx-8bit
Ministral 3 3B Instruct 251220tok/sQ8_032tok/sQ8_0
LFM2.5 24B A2B17tok/sQ5_K_M27tok/sQ6_K
Muse Glimmer 30B14tok/sQ6_K22tok/sQ6_K
DeepSeek R1 Distill Llama 8B14tok/sQ8_022tok/sQ8_0
Ministral 3 8B Reasoning 251213tok/sQ8_020tok/sQ8_0
Qwen3.8 27B OBLITERATED12tok/smlx-4bit19tok/smlx-4bit
Ministral 3 14B Reasoning 251211tok/sQ5_K_M18tok/sQ5_K_M
Qwen 3.5 122B-A10Boffloadoffload
Qwen3 Coder Nextoffloadoffload
Nemotron 3 Super 120B-A12Boffloadoffload
Mistral Small 4 119Boffloadoffload
Qwen3 Next 80B-A3Boffloadoffload
Devstral Small 2 24Boffloadoffload
Qwen3 32Boffloadoffload
Mistral Small 3.2 24Boffloadoffload
DeepSeek R1 Distill Qwen 32Boffloadoffload
Phi-4 14Boffloadoffload
DeepSeek R1 Distill Qwen 14Boffloadoffload
Devstral Small 2507offloadoffload
Magistral Small 2506offloadoffload
Magistral Small 2507offloadoffload
Magistral Small 2509offloadoffload
granite 4.2 30boffloadoffload
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
gpt-oss 120Bwon't runwon't run
GLM 4.5 Airwon't runwon't run
Llama 3.3 70Bwon't runwon't run
DeepSeek R1won't runwon't run
DeepSeek R1 0528won't runwon't run
DeepSeek R1 Distill Llama 70Bwon'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 AMD Radeon AI PRO R9700 or the NVIDIA RTX PRO 4500 Blackwell?
Choose the NVIDIA RTX PRO 4500 Blackwell for speed.
Which can run bigger models, the AMD Radeon AI PRO R9700 or the NVIDIA RTX PRO 4500 Blackwell?
At 128k context both run the same 50 models we track fully in memory.
Which is faster for local AI, the AMD Radeon AI PRO R9700 or the NVIDIA RTX PRO 4500 Blackwell?
On the models both run fully in memory, the NVIDIA RTX PRO 4500 Blackwell is about 1.6× as fast as the AMD Radeon AI PRO R9700 (estimated, median across models).

More comparisons

More hardware for local AI