Mac mini M4 32GB
32 GB unified · 120 GB/s
Prices
Every retailer we track in the US, cheapest first. Buy links may earn us a commission.
No retailer in this market lists it right now.
What it runs
Every model in the catalogue, with the best quant for this hardware. Estimates, not measurements.
| Model | Quant | Fits | Memory | Max context | Decode @8k | Prefill | Good for |
|---|---|---|---|---|---|---|---|
| Qwen 3.8 27B Alibaba Qwen | UD-Q5_K_M | fits | 19.7 GB | 31k54k with q8 cache | 4tok/sest. | 26 tok/s | – |
| Qwen 3.6 27B Alibaba Qwen | Q5_K_M | fits | 19.5 GB | 34k61k with q8 cache | 4tok/sest. | 26 tok/s | – |
| Gemma 4 31B | Q4_K_M | fits | 19.1 GB | 33k67k with q8 cache | 4tok/sest. | 23 tok/s | – |
| Qwen 3.5 27B Alibaba Qwen | Q5_K_M | fits | 19.5 GB | 33k58k with q8 cache | 4tok/sest. | 26 tok/s | – |
| Gemma 4 26B-A4B | UD-Q5_K_M | fits | 20.7 GB | 31k55k with q8 cache | 13tok/sest. | 172 tok/s | – |
| GLM 4.7 Flash Z.ai | Q5_K_M | fits | 21.0 GB | 13k22k with q8 cache | 12tok/sest. | 126 tok/s | – |
| Devstral Small 2 24B Mistral AI | Q6_K | fits | 19.9 GB | 16k30k with q8 cache | 4tok/sest. | 30 tok/s | – |
| Qwen3 Coder 30B-A3B Alibaba Qwen | Q4_K_M | fits | 18.7 GB | 34k61k with q8 cache | 18tok/sest. | 163 tok/s | chat |
| gpt-oss 20B OpenAI | MXFP4 | fits | 12.1 GB | 128k | 18tok/sest. | 163 tok/s | chat |
| Qwen3 32B Alibaba Qwen | Q4_K_M | fits | 21.1 GB | 8k16k with q8 cache | 4tok/sest. | 20 tok/s | – |
| Gemma 4 12B | Q8_0 | fits | 12.9 GB | 256k | 6tok/sest. | 60 tok/s | – |
| Qwen 3.5 9B Alibaba Qwen | Q8_0 | fits | 9.8 GB | 256k | 8tok/sest. | 76 tok/s | – |
| Gemma 3 27B | Q5_K_M | fits | 19.6 GB | 27k53k with q8 cache | 4tok/sest. | 27 tok/s | – |
| Mistral Small 3.2 24B Mistral AI | Q6_K | fits | 19.9 GB | 16k30k with q8 cache | 4tok/sest. | 30 tok/s | – |
| DeepSeek R1 Distill Qwen 32B DeepSeek | Q4_K_M | fits | 21.1 GB | 8k15k with q8 cache | 4tok/sest. | 21 tok/s | – |
| Ministral 3 14B Mistral AI | Q8_0 | fits | 15.3 GB | 45k82k with q8 cache | 5tok/sest. | 49 tok/s | – |
| Qwen3 14B Alibaba Qwen | Q8_0 | fits | 16.5 GB | 37k40k with q8 cache | 5tok/sest. | 46 tok/s | – |
| DeepSeek R1 0528 Qwen3 8B DeepSeek | Q8_0 | fits | 9.9 GB | 85k128k with q8 cache | 8tok/sest. | 80 tok/s | – |
| Qwen3 8B Alibaba Qwen | Q8_0 | fits | 9.9 GB | 40k | 8tok/sest. | 80 tok/s | – |
| Phi-4 14B Microsoft | Q8_0 | fits | 16.7 GB | 16k | 5tok/sest. | 46 tok/s | – |
| Qwen 3.5 4B Alibaba Qwen | Q8_0 | fits | 5.1 GB | 256k | 16tok/sest. | 153 tok/s | chat |
| Ministral 3 8B Mistral AI | Q8_0 | fits | 10.1 GB | 88k160k with q8 cache | 8tok/sest. | 75 tok/s | – |
| Granite 4.2 8B IBM | Q8_0 | fits | 10.6 GB | 74k128k with q8 cache | 7tok/sest. | 74 tok/s | – |
| Gemma 4 E4B | Q8_0 | fits | 8.5 GB | 128k | 9tok/sest. | 92 tok/s | – |
| LFM2.5 8B-A1B Liquid AI | Q8_0 | fits | 9.4 GB | 125k | 32tok/sest. | 536 tok/s | chat |
| Llama 3.1 8B Meta | Q6_K | fits | 7.8 GB | 111k128k with q8 cache | 10tok/sest. | 83 tok/s | – |
| DeepSeek V4 Flash DeepSeek | UD-Q2_K_XL | won't run | 91.5 GB | – | – | – | – |
| GLM 5.3 Flash Z.ai | UD-Q2_K_XL | won't run | 102.1 GB | – | – | – | – |
| Qwen 3.8 Flash Next Alibaba Qwen | UD-Q2_K_XL | won't run | 74.4 GB | – | – | – | – |
| MiniMax M2.7 MiniMax | UD-Q2_K_XL | won't run | 72.7 GB | – | – | – | – |
| Qwen 3.5 122B-A10B Alibaba Qwen | UD-Q2_K_XL | won't run | 39.9 GB | – | – | – | – |
| Nemotron 3 Super 120B-A12B NVIDIA | UD-Q2_K_XL | won't run | 51.8 GB | – | – | – | – |
| Qwen3 Coder Next Alibaba Qwen | Q4_K_M | won't run | 46.1 GB | – | – | – | – |
| gpt-oss 120B OpenAI | MXFP4 | won't run | 60.0 GB | – | – | – | – |
| Mistral Small 4 119B Mistral AI | UD-Q2_K_XL | won't run | 38.3 GB | – | – | – | – |
| Qwen 3.6 35B-A3B Alibaba Qwen | UD-Q4_K_M | won't run | 21.5 GB | 1k | – | – | – |
| GLM 4.5 Air Z.ai | Q2_K | won't run | 44.3 GB | – | – | – | – |
| Qwen 3.5 35B-A3B Alibaba Qwen | Q4_K_M | won't run | 21.4 GB | 5k8k with q8 cache | – | – | – |
| Qwen3 Next 80B-A3B Alibaba Qwen | Q4_K_M | won't run | 46.1 GB | – | – | – | – |
| Nemotron 3.5 Lightning 30B-A3B NVIDIA | UD-Q6_K_XL | won't run | 33.3 GB | – | – | – | – |
| Kimi Linear 48B-A3B Moonshot AI | Q4_K_M | won't run | 28.7 GB | – | – | – | – |
| Llama 3.3 70B Meta | Q4_K_M | won't run | 42.7 GB | – | – | – | – |
Specifications
From the manufacturer's published specifications; sources are linked where we have them.
- Chip
- Apple M4 (10-core GPU)
- Memory bus
- 128-bit
- Memory bandwidth
- 120 GB/s
- FP16 compute
- 3.76 TFLOPS
- Software
- Metal
- Unified memory
- 32 GB
- GPU can use by default
- 21.3 GB
- GPU can use, tuned
- 24 GB
- Operating system
- macos
- Processor
- M4 10-core
- Storage
- 512 GB