General model by baidu · 29.66B parameters
ERNIE 4.5 VL 28B A3B Thinking can run but will be tight at 82% of RAM. Consider a lower quantization or closing other apps.
~10 tok/s
Q4_K_M · 19.6 GB
Alternative: Q3_K_M (16.7 GB, ~12 tok/s)
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 | 33.6 GB | ~5 tok/s | ✗ |
| Q6_K | 26.2 GB | ~7 tok/s | ✗ |
| Q5_K_M | 22.6 GB | ~8 tok/s | ✗ |
| Q4_K_M Recommended | 19.6 GB | ~10 tok/s | ✓ |
| Q3_K_M | 16.7 GB | ~12 tok/s | ✓ |
| Q2_K | 13.4 GB | ~15 tok/s | ✓ |
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper — Free