publications

2026

1. arXiv

   ProjDevBench: Benchmarking AI Coding Agents on End-to-End Project Development

   Pengrui Lu^*, Shiqi Zhang^*, Yunzhong Hou^*, and 8 more authors

   arXiv preprint arXiv:2602.01655, 2026

   Abs arXiv Bib HTML Code

   An end-to-end benchmark that evaluates AI coding agents on complete project development, combining Online Judge testing with LLM-assisted code review across 20 programming problems in 8 categories.

   @article{lu2026projdevbench,
     title = {ProjDevBench: Benchmarking AI Coding Agents on End-to-End Project Development},
     author = {Lu, Pengrui and Zhang, Shiqi and Hou, Yunzhong and Ye, Lyumanshan and Huang, Chaoyi and Chen, Zixi and Zeng, Ji and Jiang, Hantao and Liu, Pengfei and Wang, Yiwei and Yang, Ming-Hsuan},
     year = {2026},
     journal = {arXiv preprint arXiv:2602.01655},
   }

2025

1. arXiv

   InnovatorBench: Evaluating Agents’ Ability to Conduct Innovative LLM Research

   Y. Wu, D. Fu, W. Si, and 9 more authors

   arXiv preprint arXiv:2510.27598, 2025

   Abs arXiv Bib

   Evaluating agents’ ability to conduct innovative LLM research.

   @article{wu2025innovatorbench,
     title = {InnovatorBench: Evaluating Agents' Ability to Conduct Innovative LLM Research},
     author = {Wu, Y. and Fu, D. and Si, W. and Huang, Z. and Jiang, M. and Li, K. and Xia, S. and Sun, J. and Xu, T. and Hu, X. and Lu, Pengrui and others},
     year = {2025},
     journal = {arXiv preprint arXiv:2510.27598},
   }

2. arXiv

   Interaction as Intelligence Part II: Asynchronous Human-Agent Rollout for Long-Horizon Task Training

   D. Fu, Y. Wu, X. Cai, and 9 more authors

   arXiv preprint arXiv:2510.27630, 2025

   Abs arXiv Bib

   Asynchronous human-agent rollout for long-horizon task training.

   @article{fu2025interaction2,
     title = {Interaction as Intelligence Part II: Asynchronous Human-Agent Rollout for Long-Horizon Task Training},
     author = {Fu, D. and Wu, Y. and Cai, X. and Ye, L. and Xia, S. and Huang, Z. and Si, W. and Xu, T. and Sun, J. and Li, K. and Jiang, M. and others},
     year = {2025},
     journal = {arXiv preprint arXiv:2510.27630},
   }

3. arXiv

   Interaction as Intelligence: Deep Research With Human-AI Partnership

   L. Ye, X. Cai, X. Wang, and 8 more authors

   arXiv preprint arXiv:2507.15759, 2025

   Abs arXiv Bib

   Deep research with human-AI partnership.

   @article{ye2025interaction,
     title = {Interaction as Intelligence: Deep Research With Human-AI Partnership},
     author = {Ye, L. and Cai, X. and Wang, X. and Wang, J. and Hu, X. and Su, J. and Nan, Y. and Wang, S. and Zhang, B. and Fan, X. and others},
     year = {2025},
     journal = {arXiv preprint arXiv:2507.15759},
   }

4. arXiv

   ParaCook: On Time-Efficient Planning for Multi-Agent Systems

   S. Zhang, X. Ma, Y. Xu, and 7 more authors

   arXiv preprint arXiv:2510.11608, 2025

   Abs arXiv Bib

   On time-efficient planning for multi-agent systems.

   @article{zhang2025paracook,
     title = {ParaCook: On Time-Efficient Planning for Multi-Agent Systems},
     author = {Zhang, S. and Ma, X. and Xu, Y. and Cao, Z. and Lu, Pengrui and Yuan, H. and Shen, T. and Zhang, Z. and Zhao, H. and others},
     year = {2025},
     journal = {arXiv preprint arXiv:2510.11608},
   }

5. arXiv

   ResearcherBench: Evaluating Deep AI Research Systems on the Frontiers of Scientific Inquiry

   Tianze Xu^*, Pengrui Lu^*, Lyumanshan Ye, and 2 more authors

   arXiv preprint arXiv:2507.16280, 2025

   Abs arXiv Bib HTML Code

   The first benchmark focused on evaluating the capabilities of Deep AI Research Systems (DARS) on frontier AI scientific questions, featuring 65 expertly curated research questions across 35 distinct AI research subjects with a dual assessment framework.

   @article{lu2025researcherbench,
     title = {ResearcherBench: Evaluating Deep AI Research Systems on the Frontiers of Scientific Inquiry},
     author = {Xu, Tianze and Lu, Pengrui and Ye, Lyumanshan and Hu, Xiangkun and Liu, Pengfei},
     year = {2025},
     journal = {arXiv preprint arXiv:2507.16280},
   }

6. arXiv

   DeepResearcher: Scaling Deep Research via Reinforcement Learning in Real-world Environments

   Yuxiang Zheng, Dayuan Fu, Xiangkun Hu, and 4 more authors

   arXiv preprint arXiv:2504.03160, 2025

   Abs arXiv Bib HTML Code

   The first comprehensive framework for end-to-end training of LLM-based deep research agents through scaling reinforcement learning in real-world environments with authentic web search interactions.

   @article{zheng2025deepresearcher,
     title = {DeepResearcher: Scaling Deep Research via Reinforcement Learning in Real-world Environments},
     author = {Zheng, Yuxiang and Fu, Dayuan and Hu, Xiangkun and Cai, Xiaojie and Ye, Lyumanshan and Lu, Pengrui and Liu, Pengfei},
     year = {2025},
     journal = {arXiv preprint arXiv:2504.03160},
   }

7. arXiv

   Deep Cognition: A Multi-Agent Framework for Collaborative Research with Real-Time Cognitive Oversight

   L. Ye, X. Cai, X. Wang, and 8 more authors

   arXiv preprint, 2025

   Abs Bib

   A multi-agent framework for collaborative research with real-time cognitive oversight.

   @article{ye2025deepcognition,
     title = {Deep Cognition: A Multi-Agent Framework for Collaborative Research with Real-Time Cognitive Oversight},
     author = {Ye, L. and Cai, X. and Wang, X. and Wang, J. and Hu, X. and Su, J. and Nan, Y. and Luo, J. and Wang, S. and Fan, X. and others},
     year = {2025},
     journal = {arXiv preprint},
   }