>>75647220I feel it is you who does not know what you are talking about, anyone who has tried LLMs can tell which models are better than another simply by using them, whether it is greater consistency or adherence to the prompt or coding.
>point to a single indication that he has "only" 2pc.picrel