Comment by stub_out

Curious how the M6 neural engine compares for local LLM inference. That's the real benchmark now.