MacBook Pro M3 Pro 18GBBudget pick
C109.5 tok/s decode
~$1,999 MSRP
Can it run?
Runs well
Using Q4_K_M in Ollama
Fit status
Runs well
Decode
218.9 tok/s
TTFT
885 ms
Safe context
46K
Memory
4.2 GB / 12.0 GB
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Agentic Coding | C | Runs well | 218.9 tok/s | 1287 ms | 92K |
| Chat | C | Runs well | 218.9 tok/s | 482 ms | 23K |
| Coding | C | Runs well | 218.9 tok/s | 885 ms | 46K |
| RAG | C | Runs well | 218.9 tok/s | 1608 ms | 92K |
| Reasoning | C | Runs well | 218.9 tok/s | 1045 ms | 46K |
How logos16v2 stablelm2 1.6b i1 (1.600000023841858B params) fits at each quantization level on Intel Arc B580 12GB (12.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 0.6 GB | Low | D30 |
Q3_K_S | 3 | 0.8 GB | Low | D30 |
NVFP4 | 4 |
Upgrade options
~$1,999 MSRP
0.9 GB |
| Medium |
| D30 |
Q4_K_M | 4 | 1.0 GB | Medium | D31 |
Q5_K_M | 5 | 1.2 GB | High | D31 |
Q6_K | 6 | 1.3 GB | High | D31 |
Q8_0 | 8 | 1.7 GB | Very High | D32 |
F16Best for your GPU | 16 | 3.3 GB | Maximum | D35 |