Can I Run Qwen 2.5 Family on Mac mini (M4 Pro, 64 GB)?

Superseded model. Qwen 2.5 Family has been superseded by Qwen 3. This page is kept for reference; the newer family is a better starting point. View Qwen 3 →

Written by Jakub Rusinowski · Last updated November 12, 2024

Yes — comfortably

Yes, comfortably — Qwen 2.5 72B Instruct at Q2_K needs about 27.2 GB of the 48 GB usable on Mac mini (M4 Pro, 64 GB), leaving ~20.8 GB spare and running at ~5.9 tok/s (estimated), with room for about 65,536 tokens of context.

Confidence: high · Recommended quantization: Q2_K · Estimated speed: ~5.9 tok/s

Mac mini (M4 Pro, 64 GB) — what it gives a model

Usable memory for models	48 GB
Memory bandwidth	273 GB/s
Form factor	Mini PC
Operating system	macOS
Memory upgradeable	No — soldered

Qwen 2.5 Family on Mac mini (M4 Pro, 64 GB): memory by quantization

Quant	Memory needed	Fits 48 GB?	Max context	Est. speed	Download
F16	147.5 GB	✗ No	—	—	144 GB
Q8_0	80 GB	✗ No	—	—	76.5 GB
Q6_K	62.5 GB	✗ No	—	—	59 GB
Q5_K_M	54.5 GB	✗ No	—	—	51 GB
Q4_K_M	47 GB	✓ Yes	8K	~3.3 tok/s	43.5 GB
Q3_K_M	34.2 GB	✓ Yes	32K	~4.6 tok/s	30.7 GB
Q2_K	27.2 GB	✓ Yes	64K	~5.9 tok/s	23.7 GB

Which Qwen 2.5 Family sizes fit

Variant	Needs at Q4_K_M	Fits?	Est. speed
Qwen 2.5 72B Instruct	47 GB	✓ Fits	~3.3 tok/s
Qwen 2.5 Coder 32B	22.3 GB	✓ Fits	~7.2 tok/s
Qwen 2.5 14B Instruct	10.9 GB	✓ Fits	~15.4 tok/s
Qwen 2.5 7B Instruct	5.9 GB	✓ Fits	~28.4 tok/s

What to watch out for

Q2_K is the only quantization that fits, and it is a heavily degraded one — expect noticeably worse output than the same model at Q4_K_M. A smaller model at Q4 is usually the better trade.
Memory on this machine is not upgradeable, so the configuration you buy is the ceiling for every model you will ever run on it.

Mac mini M4 Pro 64 GB limitations

M4 Pro memory bandwidth is half the M4 Max's, so large models that fit will still generate roughly half as fast.

Recommended setup

Ollama or LM Studio (Metal) — MLX for the fastest Apple-native throughput

How these numbers are calculated

Memory need = quantized weights + KV cache at 8,192 tokens (f16) + 0.8 GB runtime overhead.
64 GB unified memory at 273 GB/s, shared between CPU and GPU.
macOS reserves a share of unified memory for the system, so not all of it is available to a model.
Throughput is a memory-bandwidth roofline estimate, not a measurement. It is labelled "estimated" everywhere it appears.
KV cache is computed from this model's published attention configuration.

FAQ