Xiuyu Li
@sheriyuo
Researcher @StepFun_ai | Working on long-horizon tasks | Prev @RUC1937 | Opinions are my own
参加 February 2026
1.7K フォロー中 12.4K ファン
So it really just means that V4 did not perform well during training from the start 🧐
There are two hypotheses for the DeepSeek-V4's strange performance (as in, V4-Flash is about as good as we expected, but V4-Pro is disappointing given its scale):
1) failed pretrain
2) big difference in the RL/MOPD stage
Flash probably got multiple such iterations
もっと見る
