can it run
Yes, but close to the edge. 16GB against a 12GB working set: it runs, with little room for long context windows or other processes.
weights 9.1GB + KV cache 0.78GB at 4k context vs ~16GB usable → full. 9.9 GB needed of 16 GB usable — headroom for context
computed roofline: 160.4 GB/s × 0.28–0.50 efficiency window / 9.1GB (Q4_K_M) — bands, never points.
| model file | 9.1GB (GGUF Q4_K_M) |
| minimum memory | 12GB working set |
| this machine | 16GB GDDR5 (2×8 GB), 160 GB/s |
| params | 14B |
| license | mit |
| price | $35 |
easy run: ollama pull phi4:14b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ Tesla M40 (24GB, used) — cheapest machine that runs it ($55)
→ ASUS ExpertCenter Pro ET900N G3 (GB300) — this model on the biggest machine
→ every model the Tesla M60 (16GB, used) can run — full list
→ tesla-p100-16gb — the next step up from this machine's GPU