General model by davidau · 21.27B parameters
Everything this chip runs, machine by machine: MacBook Pro 14" M5 Pro, MacBook Pro 16" M5 Pro
Qwen3.5 21B Claude 4.6 Opus Deckard Heretic Uncensored Thinking fits comfortably in your 24 GB Mac using Q6_K quantization, using 79% of your RAM.
~10 tok/s
Q6_K · 18.9 GB · Estimated
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 | 24.2 GB | ~8 tok/s | ✗ |
| Q6_K Recommended | 18.9 GB | ~10 tok/s | ✓ |
| Q5_K_M | 16.4 GB | ~12 tok/s | ✓ |
| Q4_K_M | 14.2 GB | ~14 tok/s | ✓ |
| Q3_K_M | 12.1 GB | ~17 tok/s | ✓ |
| Q2_K | 9.8 GB | ~21 tok/s | ✓ |
GGUF Sources
Closest matches by use case and parameter count.
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper — Free