can it run
Comfortably. 128GB of memory against a 12GB working set leaves headroom for context, the OS, and anything else you're running.
weights 9.3GB + KV cache 0.62GB at 4k context vs ~126GB usable → full. 9.9 GB needed of 126 GB usable — headroom for context
computed roofline: 256 GB/s × 0.60–0.85 efficiency window / 9.3GB (Q4_K_M) — bands, never points.
| model file | 9.3GB (GGUF Q4_K_M) |
| minimum memory | 12GB working set |
| this machine | 128GB LPDDR5x soldered |
| params | 14B |
| license | apache-2.0 |
| price | $3449 |
easy run: ollama pull qwen3:14b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ ASUS ExpertCenter Pro ET900N G3 (GB300) — this model on the biggest machine