can it run
Comfortably. 48GB 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 ~48GB usable → full. 18.9 GB needed of 48 GB usable — headroom for context
computed roofline: 864 GB/s × 0.46–0.76 efficiency window / 17.0GB (Q4_K_M) — bands, never points.
| model file | 17GB (GGUF Q4_K_M) |
| minimum memory | 22GB working set |
| this machine | 48GB GDDR6, 864 GB/s |
| params | 27B |
| license | gemma |
| price | $7450 |
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
→ rtx-pro-6000 — the next step up from this machine's GPU
NVIDIA (used market) ↗ · Amazon ↗