It is based on Mistral which llama.cpp supports, so I assume it does run (you might need to convert to GGUF format and quantize it).