A video I found covering the improvements

https://www.youtube.com/watch?v=n_ggLjIgRcM

Will this perform better than lmstudio-community/Qwen3.8-27B-MLX-4bit on my m5 max 48gb memory mac?

undefined