can it run
Comfortably. 24GB of memory against a 10GB working set leaves headroom for context, the OS, and anything else you're running.
| model file | 8.0GB (GGUF Q4_K_M) |
| minimum memory | 10GB working set |
| this machine | 24GB GDDR5, 347 GB/s |
| params | 13B |
| license | apache-2.0 |
| price | $290 |
easy run: ollama pull llava:13b
runs with: ollama · llama.cpp · koboldcpp · lm studio
→ every model the Tesla P40 (24GB, used) can run — full list