Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
228.2 tok/s
TTFT
848 ms
Safe context
92K
Memory
32.1 GB / 184.3 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 228.2 tok/s | 1234 ms | 181K |
| Chat | C | Runs well | 228.2 tok/s | 463 ms | 46K |
| Coding | C | Runs well | 228.2 tok/s | 848 ms | 92K |
| RAG | C | Runs well | 228.2 tok/s | 1542 ms | 181K |
| Reasoning | C | Runs well | 228.2 tok/s | 1002 ms | 92K |
How gemma 3 4b it (4B params) fits at each quantization level on Mac Studio M3 Ultra 256GB (184.3 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 1.6 GB | Low | D30 |
Q3_K_S | 3 | 2.0 GB | Low | D30 |
NVFP4 | 4 | 2.2 GB | Medium | D30 |
Q4_K_M | 4 | 2.4 GB | Medium | D30 |
Q5_K_M | 5 | 2.9 GB | High | D30 |
Q6_K | 6 | 3.3 GB | High | D30 |
Q8_0 | 8 | 4.3 GB | Very High | D30 |
F16Best for your GPU | 16 | 8.2 GB | Maximum | D30 |
