can it run
Comfortably. 64GB of memory against a 24GB working set leaves headroom for context, the OS, and anything else you're running.
| model file | 18.6GB (GGUF Q4_K_M) |
| minimum memory | 24GB working set |
| this machine | up to 64GB unified |
| params | 30B-A3B MoE |
| license | apache-2.0 |
| price | $1669 |
easy run: ollama pull qwen3:30b-a3b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ ASUS ExpertCenter Pro ET900N G3 (GB300) — this model on the biggest machine
Apple ↗ · B&H Photo ↗ · Amazon ↗