Comment by speedgoose

The M7 will the much better for local LLM inference. If you can afford the memory.