can it run
| model file | 3.3GB (GGUF Q4_K_M) |
| minimum memory | 5GB working set |
| this machine | up to 4x dual-slot GPUs |
| params | 4B |
| license | gemma |
| price | $22559.99 |
easy run: ollama pull gemma3:4b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ Tesla M60 (16GB, used) — cheapest machine that runs it ($35)