Multimodal model by davidau · 19.02B parameters
Everything this chip runs, machine by machine: MacBook Pro 14" M5 Pro, MacBook Pro 16" M5 Pro
gemma 4 19B A4B it The DECKARD Heretic Uncensored Thinking fits comfortably in your 24 GB Mac using Q6_K quantization, using 71% of your RAM.
— tok/s
Q6_K · 17.0 GB · Benchmark needed
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 | 21.7 GB | ~63 tok/s | ✗ |
| Q6_K Recommended | 17.0 GB | ~83 tok/s | ✓ |
| Q5_K_M | 14.7 GB | ~97 tok/s | ✓ |
| Q4_K_M | 12.8 GB | ~114 tok/s | ✓ |
| Q3_K_M | 10.9 GB | ~138 tok/s | ✓ |
| Q2_K | 8.8 GB | ~178 tok/s | ✓ |
Affiliate disclosure: the shop icons in this table are Amazon affiliate links. As an Amazon Associate, ModelPiper earns from qualifying purchases. Every fit rating, quantization and tok/s figure below is computed from the hardware specs and is not influenced by that.
Closest matches by use case and parameter count.
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper free