Chat model by Microsoft · 7.65B parameters
Phi mini MoE instruct fits comfortably in your 24 GB Mac using Q8_0 quantization, using 38% of your RAM.
~125 tok/s
Q8_0 · 9.0 GB
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 Recommended | 9.0 GB | ~125 tok/s | ✓ |
| Q6_K | 7.1 GB | ~164 tok/s | ✓ |
| Q5_K_M | 6.2 GB | ~192 tok/s | ✓ |
| Q4_K_M | 5.4 GB | ~226 tok/s | ✓ |
| Q3_K_M | 4.7 GB | ~273 tok/s | ✓ |
| Q2_K | 3.8 GB | ~354 tok/s | ✓ |
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper — Free