Question 1

Can a Mac mini M6 (24GB) run Qwen3 Coder 30B-A3B (MoE)?

Accepted Answer

Yes — Qwen3 Coder 30B-A3B (MoE) runs at Q2_K (11.3 GB) at about 53–79 tok/s on a Mac mini M6 (24GB).

Question 2

How much memory does Qwen3 Coder 30B-A3B (MoE) need?

Accepted Answer

Qwen3 Coder 30B-A3B (MoE) needs about 12.5 GB at 4K context at its smallest quantization. A Mac mini M6 (24GB) has 24.0 GB.

Quant	Size	Speed	Fit
Q2_K	11.3 GB	53–79 tok/s	Runs great
Q3_K_M	14.7 GB	46–68 tok/s	Tight / slow
Q4_K_M	18.6 GB	40–59 tok/s	Won't fit
Q5_K_M	21.7 GB	36–54 tok/s	Won't fit
Q6_K	25.1 GB	32–49 tok/s	Won't fit
Q8_0	32.5 GB	27–40 tok/s	Won't fit
FP16	61.1 GB	16–24 tok/s	Won't fit