Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
64.6 tok/s
TTFT
2995 ms
Safe context
42K
Memory
13.3 GB / 34.6 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 62.7 tok/s | 4494 ms | 75K |
| Chat | C | Runs well | 62.7 tok/s | 1685 ms | 22K |
| Coding | C | Runs well | 64.6 tok/s | 2995 ms | 42K |
| RAG | C | Runs well | 62.7 tok/s | 5618 ms | 75K |
| Reasoning | C | Runs well | 62.7 tok/s | 3652 ms | 42K |
How Qwen3.5 9B (9B params) fits at each quantization level on MacBook Pro M4 Max 48GB (34.6 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 3.5 GB | Low | D32 |
Q3_K_S | 3 | 4.4 GB | Low | D32 |
NVFP4 | 4 |
huggingface-cli download hf-lmstudio-community--qwen3-5-9b-gguf5.0 GB |
| Medium |
| D32 |
Q4_K_M | 4 | 5.5 GB | Medium | D33 |
Q5_K_M | 5 | 6.5 GB | High | D33 |
Q6_K | 6 | 7.4 GB | High | D34 |
Q8_0 | 8 | 9.6 GB | Very High | D35 |
F16Best for your GPU | 16 | 18.5 GB | Maximum | C41 |