General model by jsfs11 · 24.15B parameters
Everything this chip runs, machine by machine: MacBook Pro 14" M5 Pro, MacBook Pro 16" M5 Pro
MixtureofMerges MoE 4x7b v4 can run but will be tight at 77% of RAM. Consider a lower quantization or closing other apps.
— tok/s
Q5_K_M · 18.5 GB · Benchmark needed
Alternative: Q4_K_M (16.1 GB, ~23 tok/s)
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 | 27.4 GB | ~13 tok/s | ✗ |
| Q6_K | 21.4 GB | ~17 tok/s | ✗ |
| Q5_K_M Recommended | 18.5 GB | ~20 tok/s | ✓ |
| Q4_K_M | 16.1 GB | ~23 tok/s | ✓ |
| Q3_K_M | 13.7 GB | ~28 tok/s | ✓ |
| Q2_K | 11.0 GB | ~36 tok/s | ✓ |
Closest matches by use case and parameter count.
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper — Free