What can a Mac mini M4 16GB run?
16GB · 120 GB/s · 65W · $599
The gateway drug: a $599 16 GB mini that quietly runs 8B Q4 models at chat speed on 65W.
Runs great14
Comfortable headroom — daily-driver territory.
| Model | Quant | Est. speed | Max context |
|---|---|---|---|
| deepseek moe 16b chat | FP16 | Infinity tok/s | — |
| qwen2 7b instruct | FP16 | Infinity tok/s | — |
| Qwen3-0.6B | FP16 | 60 tok/s | 33k |
| Gemma-3n-E4B | FP16 | 27 tok/s | 40k |
| Llama-3.2-3B | Q8_0 | 22 tok/s | 93k |
| SmolLM2-1.7B | FP16 | 22 tok/s | 8k |
| Phi-4-mini | Q8_0 | 18 tok/s | 81k |
| Gemma-3-4B | Q8_0 | 18 tok/s | 81k |
| Qwen3-4B | Q8_0 | 18 tok/s | 33k |
| Mistral-7B-Instruct | Q4_K_M | 17 tok/s | 33k |
| DeepSeek-R1-Distill-Qwen-7B | Q4_K_M | 16 tok/s | 60k |
| Qwen2.5-Coder-7B | Q4_K_M | 16 tok/s | 60k |
| Llama-3.1-8B | Q4_K_M | 15 tok/s | 57k |
| Qwen3-8B | Q4_K_M | 15 tok/s | 33k |
Tight squeeze13
It fits, but close the browser tabs and mind the context.
| Model | Quant | Est. speed | Max context |
|---|---|---|---|
| Qwen3-30B-A3B | IQ2_M | 67 tok/s | 7k |
| Qwen3-Coder-30B-A3B | IQ2_M | 67 tok/s | 7k |
| GPT-OSS-20B | Q3_K_M | 42 tok/s | 8k |
| Mistral-Small-24B | IQ2_M | 9 tok/s | 21k |
| Qwen3.8 | IQ2_M | 8 tok/s | 15k |
| Qwen3.8 27B | IQ2_M | 8 tok/s | 15k |
| Gemma-3-27B | IQ2_M | 8 tok/s | 15k |
| qwen35 27B | IQ2_M | 8 tok/s | 14k |
| Qwen3-14B | Q4_K_M | 7 tok/s | 8k |
| Phi-4-14B | Q5_K_M | 7 tok/s | 7k |
| DeepSeek-R1-Distill-Qwen-14B | Q5_K_M | 7 tok/s | 6k |
| Gemma-3-12B | Q6_K | 7 tok/s | 6k |
| Codestral-22B | Q3_K_M | 7 tok/s | 4k |
Measured on this card0
Everything here was benchmarked on this exact card by someone who owns one — the tables above are our arithmetic; these numbers came off real machines.
Nobody has measured anything on this card yet. Paste a benchmark — it takes one keystroke and no form.