can it run
Comfortably. 64GB of memory against a 21GB working set leaves headroom for context, the OS, and anything else you're running.
| model file | 16.5GB (GGUF Q4_K_M) |
| minimum memory | 21GB working set |
| this machine | 64GB LPDDR5, 204.8 GB/s |
| params | 27B |
| license | apache-2.0 |
| price | $3499 |
easy run: ollama pull qwen3.8:27b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ ASUS ExpertCenter Pro ET900N G3 (GB300) — this model on the biggest machine