can it run
Yes, but close to the edge. 8GB against a 8GB working set: it runs, with little room for long context windows or other processes.
weights 6.2GB + KV cache 2GB at 4k context vs ~8GB usable → tight. 8.2 of 8 GB usable — barely over; real with q8 KV cache (halves KV) or shorter context
computed roofline: 192.3 GB/s × 0.28–0.50 efficiency window / 6.2GB (Q4_K_M) — bands, never points.
| model file | 6.2GB (GGUF Q4_K_M) |
| minimum memory | 8GB working set |
| this machine | 8GB GDDR5, 192 GB/s |
| params | large-v3 |
| license | apache-2.0 |
| price | $100 |
easy run: no stock runtime yet
→ Tesla M60 (16GB, used) — cheapest machine that runs it ($35)
→ every model the Tesla P4 (8GB, used) can run — full list
→ tesla-p100-16gb — the next step up from this machine's GPU