can it run
Comfortably. 96GB of memory against a 4GB working set leaves headroom for context, the OS, and anything else you're running.
| model file | 2.0GB (GGUF Q4_K_M) |
| minimum memory | 4GB working set |
| this machine | 96GB base, config to 512GB (512GB ships late Oct) |
| params | 3B |
| license | llama3.2 (gated on HF) |
| price | $9499 |
easy run: ollama pull llama3.2:3b
runs with: ollama · llama.cpp · koboldcpp · lm studio
no alternatives — every machine in the catalog can run this one.