Singularity-Bench: Can Agents Build Agents?with Yang Zhou, Krish Agarwal, Zhuoming Chen, Zerui Wang +142 versionsW_k5jdtrvav2 · current
QUEST: Training Frontier Deep Research Agents with Fully Synthetic TasksJian Xie, Tianhe Lin, Zilu Wang et al.0 citations2605.24218v22026arXiv
Does RAG Know When Retrieval Is Wrong? Diagnosing Context Compliance under Knowledge ConflictYihang Chen, Pin Qian, Su Wang et al.0 citations2605.14473v42026arXiv
How Deep Can LLMs Learn to Reason? Expressiveness Is KeyTianle Wang, Zhaoyang Wang, Guangchen Lan et al.0 citations2605.06638v42026arXiv