readygpu

Mac mini M4 Pro 48GB vs AMD Radeon AI PRO R9700

For running models locally, priced in Germany. At 32k context the first runs 57 models fully in memory, the second 57.

In short

The Mac mini M4 Pro 48GB has more memory for models: 48 GB against 32 GB. At 32k context both run the same 57 models we track fully in memory. On the models both run fully in memory, the AMD Radeon AI PRO R9700 is about 2.0× as fast as the Mac mini M4 Pro 48GB (estimated, median across models). In Germany the Mac mini M4 Pro 48GB costs 2.699 € at eBay (used) and the AMD Radeon AI PRO R9700 2.390 € at eBay: the AMD Radeon AI PRO R9700 is 310 € cheaper.

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

Mac mini M4 Pro 48GB

2.699 €

eBay, used (asking)

Add to build
AMD Radeon AI PRO R9700

2.390 €

eBay

Add to build

Specifications

SpecMac mini M4 Pro 48GBAMD Radeon AI PRO R9700
ChipApple M4 Pro (20-core GPU)–
Memory bus256-bit256-bit
Memory bandwidth273 GB/s640 GB/s
FP16 compute8.08 TFLOPS191 TFLOPS
SoftwareMetalROCm
Unified memory48 GB–
GPU can use by default36 GB–
GPU can use, tuned40 GB–
Operating systemmacos–
ProcessorM4 Pro 14-core–
Storage512 GB–
GPU–Radeon AI PRO R9700
Memory–32 GB GDDR6
Board power–300 W
PCIe generation–Gen 5
PCIe lanes–x16

What each runs

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

ModelMac mini M4 Pro 48GBAMD Radeon AI PRO R9700
Qwen 3.8 27B6tok/sQ8_011tok/sQ8_0
Qwen 3.6 27B6tok/sQ8_011tok/sQ8_0
Gemma 4 31B5tok/sQ8_012tok/sQ6_K
Qwen 3.6 35B-A3B28tok/sQ8_064tok/sQ6_K
Qwen 3.5 27B6tok/sQ8_011tok/sQ8_0
Gemma 4 26B-A4B21tok/sQ8_042tok/sQ8_0
Qwen 3.5 35B-A3B28tok/sQ8_067tok/sQ6_K
GLM 4.7 Flash15tok/sQ8_036tok/sQ6_K
Devstral Small 2 24B6tok/sQ8_012tok/sQ8_0
Nemotron 3.5 Lightning 30B-A3B30tok/sUD-Q6_K_XL68tok/sUD-Q5_K_M
Kimi Linear 48B-A3B44tok/sQ5_K_M99tok/sQ4_K_M
gpt-oss 20B33tok/sMXFP465tok/sMXFP4
Qwen3 Coder 30B-A3B17tok/sQ8_037tok/sQ6_K
Qwen3 32B5tok/sQ6_K11tok/sQ5_K_M
Gemma 4 12B13tok/sQ8_026tok/sQ8_0
Qwen 3.5 9B17tok/sQ8_033tok/sQ8_0
Mistral Small 3.2 24B6tok/sQ8_012tok/sQ8_0
Gemma 3 27B6tok/sQ8_011tok/sQ8_0
DeepSeek R1 Distill Qwen 32B5tok/sQ6_K11tok/sQ5_K_M
Ministral 3 14B9tok/sQ8_018tok/sQ8_0
Qwen3 14B8tok/sQ8_017tok/sQ8_0
DeepSeek R1 0528 Qwen3 8B13tok/sQ8_026tok/sQ8_0
Qwen3 8B13tok/sQ8_026tok/sQ8_0
Phi-4 14B8tok/sQ8_016tok/sQ8_0
Qwen 3.5 4B32tok/sQ8_063tok/sQ8_0
Ministral 3 8B13tok/sQ8_026tok/sQ8_0
Granite 4.2 8B12tok/sQ8_024tok/sQ8_0
Gemma 4 E4B19tok/sQ8_038tok/sQ8_0
LFM2.5 8B-A1B40tok/sQ8_079tok/sQ8_0
Llama 3.1 8B16tok/sQ6_K32tok/sQ6_K
Qwen2.5 0.5B190tok/sQ8_0377tok/sQ8_0
Qwen3.5 0.8B144tok/sQ8_0285tok/sQ8_0
Qwen3.5 2B96tok/sQ5_K_M190tok/sQ5_K_M
DeepSeek R1 Distill Qwen 1.5B63tok/sQ8_0124tok/sQ8_0
MiniCPM5 2B43tok/sQ8_086tok/sQ8_0
Spark X2.5 4B41tok/smlx-4bit82tok/smlx-4bit
LFM2.5 2.6B36tok/sQ8_072tok/sQ8_0
LFM2.5 VL 3B36tok/sQ8_072tok/sQ8_0
Mellum2.1 12B A2.5B Thinking34tok/sQ8_067tok/sQ8_0
granite vision 4.1 4b28tok/sQ8_056tok/sQ8_0
granite 4.2 3b27tok/sQ8_054tok/sQ8_0
Kimi VL A3B Thinking 250626tok/sQ8_051tok/sQ8_0
Kimi VL A3B25tok/smlx-8bit49tok/smlx-8bit
LFM2.5 24B A2B22tok/sQ8_044tok/sQ8_0
GLM 4 9B 041419tok/smlx-5bit37tok/smlx-5bit
DeepSeek R1 Distill Qwen 7B18tok/sQ8_035tok/sQ8_0
GLM 4.6V Flash16tok/sQ8_031tok/sQ8_0
DeepSeek R1 Distill Llama 8B14tok/sQ8_027tok/sQ8_0
GLM 4.1V 9B Thinking14tok/smlx-8bit27tok/smlx-8bit
Ministral 3 8B Reasoning 251213tok/sQ8_026tok/sQ8_0
DeepSeek R1 Distill Qwen 14B8tok/sQ8_016tok/sQ8_0
Qwen3.8 27B OBLITERATED8tok/smlx-4bit15tok/smlx-4bit
Devstral Small 25076tok/sQ8_012tok/sQ8_0
Magistral Small 25066tok/sQ8_012tok/sQ8_0
Magistral Small 25076tok/sQ8_012tok/sQ8_0
Magistral Small 25096tok/sQ8_012tok/sQ8_0
granite 4.2 30b5tok/sQ6_K12tok/sQ5_K_M
Qwen 3.5 122B-A10Bwon't runoffload
Nemotron 3 Super 120B-A12Bwon't runoffload
Qwen3 Coder Nextwon't runoffload
Mistral Small 4 119Bwon't runoffload
GLM 4.5 Airwon't runoffload
Qwen3 Next 80B-A3Bwon't runoffload
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
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 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 4.5Vwon'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 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 Mac mini M4 Pro 48GB or the AMD Radeon AI PRO R9700?
Choose the AMD Radeon AI PRO R9700 for speed. Choose the AMD Radeon AI PRO R9700 to spend less.
Which can run bigger models, the Mac mini M4 Pro 48GB or the AMD Radeon AI PRO R9700?
At 32k context both run the same 57 models we track fully in memory.
Which is faster for local AI, the Mac mini M4 Pro 48GB or the AMD Radeon AI PRO R9700?
On the models both run fully in memory, the AMD Radeon AI PRO R9700 is about 2.0× as fast as the Mac mini M4 Pro 48GB (estimated, median across models).
What is the price difference between the Mac mini M4 Pro 48GB and the AMD Radeon AI PRO R9700?
In Germany the Mac mini M4 Pro 48GB costs 2.699 € at eBay (used) and the AMD Radeon AI PRO R9700 2.390 € at eBay: the AMD Radeon AI PRO R9700 is 310 € cheaper.

More comparisons