News

Qwen3-Max arrives in preview with 1 trillion parameters, blazing fast response speed, and API availability

With more than 1 trillion parameters, Qwen3-Max-Preview signals Alibaba Cloud’s ongoing investment in scaling AI systems. As ...

Adjusting the Sequence of Training Data Makes Large Models Smarter! No Need to Expand Model/Data Scale

In this regard, Microsoft Research Asia has proposed a novel paradigm for organizing text data called DELT (Data Efficacy in ...

Adjusting the Sequence of Training Data Can Make Large Models Smarter! No Need to Expand Model/Data Scale

In this regard, Microsoft Research Asiahas proposed a brand new text data organization paradigm called DELT(Data Efficacy in LM Training). By introducing data sorting strategies, it fully taps into ...