Comment by pinyen_cheng

I think the future should be efficient quantization which can make model with the same level as Qwen 3.8 27b now accessible from our laptop.