General model by Google · 26.54B parameters · Released 2026-03-11
gemma 4 26B A4B it can run but will be tight at 85% of RAM. Consider a lower quantization or closing other apps.
~9 tok/s
Q5_K_M · 20.3 GB
Alternative: Q4_K_M (17.6 GB, ~11 tok/s)
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 | 30.1 GB | ~6 tok/s | ✗ |
| Q6_K | 23.5 GB | ~8 tok/s | ✗ |
| Q5_K_M Recommended | 20.3 GB | ~9 tok/s | ✓ |
| Q4_K_M | 17.6 GB | ~11 tok/s | ✓ |
| Q3_K_M | 15.0 GB | ~13 tok/s | ✓ |
| Q2_K | 12.1 GB | ~17 tok/s | ✓ |
GGUF Sources
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper — Free