themarkymarklast weekQwen 3.8 27b performs very favorably. I am not really sure where I would put it in ChatGPT line up, it is tiny in comparison but it is a very strong model…113 00E-8 BEE