X.com06/17, 20:20ResearchFine-Tuning Open Models Can Exceed Frontier ModelsBase @Alibaba_Qwen performs well with good prompting.LangChainfine-tuning