>>106307962
the only model worth using for productive tasks is llama 3.1 405b, the moe meme killed every other use case so there's basically been no new models for an entire year