Apple · M4 · unified memory

What runs on a Mac Mini M4 Pro 48GB

48 GB at 273 GB/s. The largest model that fits at Q4_K_M is Mixtral 8x7B, at roughly 19.1 tokens per second. Unified memory means the operating system takes a share, so budget around three quarters of the total.

Runs comfortably

6.1 GB of 36.0 GB · 17%
036 GB
Weights 4.6 GB
KV cache 1.0 GB
Runtime overhead 0.5 GB

Llama 3.1 8B at Q4_K_M leaves 29.9 GB spare on a Mac Mini M4 Pro 48GB. There is room to raise the context length or move up a quantisation level.

Generation34.1tok/s
Prompt processing445tok/s
Max context128Ktokens
KV per 1K tokens0GB

Every quantisation of Llama 3.1 8B on a Mac Mini M4 Pro 48GB

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 15.0 GB 16.5 GB Runs comfortably 128K 11.3
INT8 / W8A8 8.50 7.7 GB 9.2 GB Runs comfortably 128K 21.3
Q8_0 (GGUF) 8.50 7.7 GB 9.2 GB Runs comfortably 128K 21.3
FP8 (E4M3) 8.00 7.5 GB 9.0 GB Runs comfortably 128K 21.9
Q6_K 6.56 6.1 GB 7.7 GB Runs comfortably 128K 26.4
AWQ 4-bit 4.25 5.4 GB 6.9 GB Runs comfortably 128K 29.6
GPTQ 4-bit 4.25 5.4 GB 6.9 GB Runs comfortably 128K 29.6
MXFP4 4.25 5.4 GB 6.9 GB Runs comfortably 128K 29.6
Q5_K_M 5.67 5.3 GB 6.8 GB Runs comfortably 128K 30.1
Q5_K_S 5.52 5.2 GB 6.7 GB Runs comfortably 128K 30.9
Q4_K_M 4.85 4.6 GB 6.1 GB Runs comfortably 128K 34.1
Q4_K_S 4.58 4.4 GB 5.9 GB Runs comfortably 128K 35.6
Q4_0 4.55 4.4 GB 5.9 GB Runs comfortably 128K 35.8
IQ4_XS 4.25 4.1 GB 5.6 GB Runs comfortably 128K 37.7
Q3_K_M 3.91 3.8 GB 5.4 GB Runs comfortably 128K 40.1
IQ3_M 3.70 3.7 GB 5.2 GB Runs comfortably 128K 41.7
IQ3_XXS 3.06 3.2 GB 4.7 GB Runs comfortably 128K 47.6
Q2_K 2.63 2.8 GB 4.3 GB Runs comfortably 128K 52.5
IQ2_XXS 2.06 2.3 GB 3.9 GB Runs comfortably 128K 61.0
IQ1_M 1.75 2.1 GB 3.6 GB Runs comfortably 128K 66.8

Models on a Mac Mini M4 Pro 48GB

Q4_K_M at 8K context. Bandwidth sets the speed; capacity sets the ceiling.

ModelParamsWeightsVerdictMax ctxtok/s
DeepSeek-R1 671B-A37B 671B 379.0 GB Won't fit — 2.04
Qwen3 235B-A22B 235.1B 132.8 GB Won't fit — 3.70
gpt-oss 120B-A5.1B 116.8B 66.0 GB Won't fit — 20.4
Llama 4 Scout 109B-A17B 109B 61.7 GB Won't fit — 6.52
Qwen2.5 72B 72.7B 41.2 GB Won't fit — 2.43
Llama 3.1 70B 70.5B 40.0 GB Won't fit — 2.63
Mixtral 8x7B 46.7B 26.4 GB Runs comfortably 32K 19.1
Command R 35B 35.0B 20.1 GB Runs comfortably 12K 7.00
Yi-1.5 34B 34.4B 19.5 GB Runs comfortably 32K 8.57
Qwen3 32B 32.8B 18.6 GB Runs comfortably 67K 8.93
Qwen2.5-Coder 32B 32.8B 18.6 GB Runs comfortably 67K 8.92
DeepSeek-R1-Distill-Qwen 32B 32.8B 18.6 GB Runs comfortably 67K 8.92
Qwen3 30B-A3B 30.5B 17.3 GB Runs comfortably 128K 52.4
Gemma 3 27B 27.4B 15.6 GB Runs comfortably 128K 10.7
Mistral Small 24B 23.6B 13.4 GB Runs comfortably 32K 12.5
gpt-oss 20B-A3.6B 20.9B 11.9 GB Runs comfortably 128K 63.2
Qwen3 14B 14.8B 8.5 GB Runs comfortably 128K 19.3
Phi-4 14B 14.7B 8.4 GB Runs comfortably 16K 19.1
Mistral NeMo 12B 12.3B 7.0 GB Runs comfortably 128K 22.9
Gemma 3 12B 12.2B 7.0 GB Runs comfortably 128K 23.2
GLM-4 9B 9.4B 5.4 GB Runs comfortably 128K 31.4
Qwen3 8B 8.2B 4.7 GB Runs comfortably 128K 33.0
Llama 3.1 8B 8.0B 4.6 GB Runs comfortably 128K 34.1
DeepSeek-R1-Distill-Llama 8B 8.0B 4.6 GB Runs comfortably 128K 34.1
Qwen2.5 7B 7.6B 4.4 GB Runs comfortably 128K 37.9
Mistral 7B v0.3 7.3B 4.1 GB Runs comfortably 32K 37.8
Gemma 3 4B 4.3B 2.5 GB Runs comfortably 128K 64.6
Qwen3 4B 4.0B 2.3 GB Runs comfortably 128K 60.6
Llama 3.2 3B 3.2B 1.8 GB Runs comfortably 128K 76.1
Llama 3.2 1B 1.2B 0.7 GB Runs comfortably 128K 204

Direct answers

Llama 3.1 8B on Mac Mini M4 Pro 48GB
8.0BLlama 3.1 70B on Mac Mini M4 Pro 48GB
70.5BLlama 3.2 3B on Mac Mini M4 Pro 48GB
3.2BLlama 3.2 1B on Mac Mini M4 Pro 48GB
1.2BLlama 4 Scout 109B-A17B on Mac Mini M4 Pro 48GB
109BQwen3 8B on Mac Mini M4 Pro 48GB
8.2BQwen3 14B on Mac Mini M4 Pro 48GB
14.8BQwen3 32B on Mac Mini M4 Pro 48GB
32.8B