Kimi K3 Disappoints Users Who Favor American AI Models
Source Posts on X
chinese models are 7-12 months behind easily. training on some benchmarks is great, but you’ll start to see the tweets trickling through the roof soon enough it falls apart on real world complex problems they’ll cry. also be interested how it does on the remote labor index or arc agi 3.
looks like the results are coming in and there’s general disappointment for kimi k3 i did try warning you that china are seven months behind, stick with american models, nothing comes remotely close.