can it run
Comfortably. 192GB of memory against a 22GB working set leaves headroom for context, the OS, and anything else you're running.
weights 17GB + KV cache 1.94GB at 4k context vs ~190GB usable → full. 18.9 GB needed of 190 GB usable — headroom for context
computed roofline: 256 GB/s × 0.60–0.85 efficiency window / 17.0GB (Q4_K_M) — bands, never points.
| model file | 17GB (GGUF Q4_K_M) |
| minimum memory | 22GB working set |
| this machine | 192GB LPDDR5X-8533 unified (160GB GPU-assignable) |
| params | 27B |
| license | gemma |
| price | $None |
easy run: ollama pull gemma3:27b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ ASUS ExpertCenter Pro ET900N G3 (GB300) — this model on the biggest machine
HotHardware (announce) ↗ · Amazon ↗