Mac mini M4 64GBBudget pick
C4 tok/s decode
~$1,099 MSRP
Can it run?
Too heavy
Using Q4_K_M in Ollama
Fit status
Too heavy
Decode
9.1 tok/s
TTFT
21197 ms
Safe context
13K
Memory
30.4 GB / 24.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | F | Too heavy | 9.1 tok/s | 30832 ms | 21K |
| Chat | D | Very compromised (needs ~2.8 GB host RAM) | 8.1 tok/s | 13037 ms | 7K |
| Coding | F | Too heavy | 9.1 tok/s | 21197 ms | 13K |
| RAG | F | Too heavy | 9.1 tok/s | 38539 ms | 21K |
| Reasoning | F | Too heavy | 9.1 tok/s | 25051 ms | 13K |
How Qwen3.5 35B A3B (35B params) fits at each quantization level on NVIDIA L4 24GB (24.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 13.7 GB | Low | C42 |
Q3_K_SBest for your GPU | 3 | 17.2 GB | Low | C45 |
NVFP4 | 4 | 19.6 GB | Medium | C45 |
Q4_K_M | 4 | 21.3 GB | Medium | C45 |
Q5_K_M | 5 | 25.2 GB | High | F0 |
Q6_K | 6 | 28.7 GB | High | F0 |
Q8_0 | 8 | 37.5 GB | Very High | F0 |
F16 | 16 | 71.8 GB | Maximum | F0 |
Upgrade options
~$1,099 MSRP
~$1,599 MSRP
NVIDIA A100 40GBBiggest leap~$10,000 MSRP
RTX PRO 5000 Blackwell 48GBNVIDIA upgrade