Comment by itake

does your comment depend on the OS? I thought MLX has better performance on MacOS than llama.cpp