General model by ibm-research · 3.37B parameters
PowerMoE 3b fits comfortably in your 24 GB Mac using Q8_0 quantization, using 18% of your RAM.
~199 tok/s
Q8_0 · 4.3 GB
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 Recommended | 4.3 GB | ~199 tok/s | ✓ |
| Q6_K | 3.4 GB | ~261 tok/s | ✓ |
| Q5_K_M | 3.0 GB | ~307 tok/s | ✓ |
| Q4_K_M | 2.7 GB | ~359 tok/s | ✓ |
| Q3_K_M | 2.3 GB | ~434 tok/s | ✓ |
| Q2_K | 2.0 GB | ~564 tok/s | ✓ |
GGUF Sources
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper — Free