Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
101.4 tok/s
TTFT
1908 ms
Safe context
83K
Memory
35.7 GB / 184.3 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 101.4 tok/s | 2776 ms | 159K |
| Chat | C | Runs well | 101.4 tok/s | 1041 ms | 42K |
| Coding | C | Runs well | 101.4 tok/s | 1908 ms | 83K |
| RAG | C | Runs well | 101.4 tok/s | 3470 ms | 159K |
| Reasoning | C | Runs well | 101.4 tok/s | 2255 ms | 83K |
How Qwen3.5 9B (9B params) fits at each quantization level on Mac Studio M3 Ultra 256GB (184.3 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 3.5 GB | Low | D30 |
Q3_K_S | 3 | 4.4 GB | Low | D30 |
NVFP4 | 4 | 5.0 GB | Medium | D30 |
Q4_K_M | 4 | 5.5 GB | Medium | D30 |
Q5_K_M | 5 | 6.5 GB | High | D30 |
Q6_K | 6 | 7.4 GB | High | D30 |
Q8_0 | 8 | 9.6 GB | Very High | D31 |
F16Best for your GPU | 16 | 18.5 GB | Maximum | D32 |
huggingface-cli download hf-unsloth--qwen3-5-9b-gguf