开放式自我改进的智能体能够自主修改自身结构设计以提升能力并克服预定义架构的限制，从而减少对人类干预的依赖。我们引入了群体进化代理（GEA），一种新的开放式自我改进范式，将一组智能体视为基本的进化单位，使在进化过程中能够显式地共享和重用组内的经验。与现有开放式自我进化范式采用树状结构进化不同，GEA克服了由于孤立进化分支导致的探索多样性利用效率低下的局限。我们在具有挑战性的编码基准上评估了GEA，其在SWE-bench Verified上显著优于最先进的自我进化方法（71.0% vs. 56.7%，Polyglot上88.3% vs. 68.3%），并且在两个基准上与顶级人工设计的智能体框架相匹配或超过（71.8%和52.0%）。分析表明，GEA更有效地将早期阶段的探索多样性转化为持续的长期进步，在相同数量的进化智能体下实现更强的性能。此外，GEA在不同的编码模型中表现出一致的可迁移性，并且具有更强的鲁棒性，修复框架级bug的平均迭代次数为1.4次，而自我进化方法为5次。
英文摘要：
Open-ended self-improving agents can autonomously modify their own structural designs to advance their capabilities and overcome the limits of pre-defined architectures, thus reducing reliance on human intervention. We introduce Group-Evolving Agents (GEA), a new paradigm for open-ended self-improvements, which treats a group of agents as the fundamental evolutionary unit, enabling explicit experience sharing and reuse within the group throughout evolution. Unlike existing open-ended self-evolving paradigms that adopt tree-structured evolution, GEA overcomes the limitation of inefficient utilization of exploratory diversity caused by isolated evolutionary branches. We evaluate GEA on challenging coding benchmarks, where it significantly outperforms state-of-the-art self-evolving methods (71.0% vs. 56.7% on SWE-bench Verified, 88.3% vs. 68.3% on Polyglot) and matches or exceeds top human-designed agent frameworks (71.8% and 52.0% on two benchmarks, respectively). Analysis reveals that GEA more effectively converts early-stage exploratory diversity into sustained, long-term progress, achieving stronger performance under the same number of evolved agents. Furthermore, GEA exhibits consistent transferability across different coding models and greater robustness, fixing framework-level bugs in 1.4 iterations on average, versus 5 for self-evolving methods.
MedZERO: Self-Evolving Agents for Open-Ended Medical Reasoning Through Controlled Knowledge Accumulation
The Chinese University of Hong Kong（香港中文大学）；Zhongshan Hospital, Fudan University（复旦大学附属中山医院）；South China University of Technology（华南理工大学）；Chinese Academy of Sciences（中国科学院）
Safety Must Precede the Deployment of Open-Ended AI
CISPA-Helmholtz Center of Information Security（CISPA-海德堡信息安全中心）；MPI for Intelligent Systems, ELLIS Institute Tübingen, Tübingen AI Center（智能系统Max Planck研究所、图宾根ELLIS研究所、图宾根人工智能中心）
DIRO & Institut Courtois, Université de Montréal（DIRO与Courtois研究所，蒙特利尔大学）；Mila – Québec AI Institute（魁北克人工智能研究所）；Microsoft Research（微软研究院）；Canada CIFAR AI Chair（加拿大CIFAR人工智能主席）
DéjàQ: Open-Ended Evolution of Diverse, Learnable and Verifiable Problems
Willem Röpke AI Lab, Vrije Universiteit Brussel Belgium（维尔姆·罗普克人工智能实验室，布鲁塞尔自由大学）；FLAIR, University of Oxford United Kingdom（FLAIR，牛津大学）；University College London United Kingdom（伦敦大学学院）
CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery
MIT（麻省理工学院）；NUS（新加坡国立大学）；MiniMax；McGill（麦吉尔大学）；Stanford（斯坦福大学）；SambaNova；Meta；Singapore-MIT Alliance for Research and Technology（新加坡-麻省理工联合研究技术联盟）；Amazon（亚马逊）；Microsoft（微软）
MAGELLAN: Metacognitive predictions of learning progress guide autotelic LLM agents in large goal spaces
Department of XXX, University of YYY, Location, Country（XXX部门，YYYY大学，地点，国家）；School of ZZZ, Institute of WWW, Location, Country（ZZZ学院，WWW研究所，地点，国家）；Inria (Flowers), University of Bordeaux, France（Inria（Flowers），波尔多大学，法国）；MIT, Computational Cognitive Science Lab, Cambridge, MA, USA（麻省理工学院，计算认知科学实验室，马萨诸塞州剑桥，美国）
Collaborative Auto-Curricula Multi-Agent Reinforcement Learning with Graph Neural Network Communication Layer for Open-ended Wildfire-Management Resource Distribution
SOLAR: A Self-Optimizing Open-Ended Autonomous Agent for Lifelong Learning and Continual Adaptation
Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of Singapore, Singapore（眼科学系，Yong Loo Lin医学院，新加坡国立大学，新加坡）；Department of Computational and Data Sciences, Indian Institute of Science, Bangalore, Karnataka, India（计算与数据科学系，印度科学研究院，班加罗尔，卡纳塔克邦，印度）
When Stories Evolve: Benchmarking LLM Storytelling Across Agent Architectures in Open-Ended World Simulations
The University of Hong Kong（香港大学）；Peking University（北京大学）；Tsinghua University（清华大学）；Beijing Institute of Mathematical Sciences and Applications (BIMSA)（北京数学科学与应用研究院）
G-Zero: Self-Play for Open-Ended Generation from Zero Data
Washington University in St. Louis（华盛顿大学圣路易斯分校）；University of Virginia（弗吉尼亚大学）；University of Maryland（马里兰大学）；University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）
A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence
Princeton University（普林斯顿大学）；Princeton AI Lab（普林斯顿人工智能实验室）；Tsinghua University（清华大学）；Carnegie Mellon University（卡内基梅隆大学）；University of Sydney（悉尼大学）；Shanghai Jiao Tong University（上海交通大学）；Pennsylvania State University（宾夕法尼亚州立大学）；University of Michigan（密歇根大学）；Oregon State University（俄勒冈州立大学）；The Chinese University of Hong Kong（香港中文大学）；Fudan University（复旦大学）；The Hong Kong University of Science and Technology (Guangzhou)（香港科学与技术大学（广州））；The University of Hong Kong（香港大学）；University of California, Santa Barbara（加州大学圣芭芭拉分校）；University of California San Diego（加州大学圣地亚哥分校）；University of Edinburgh（爱丁堡大学）；University of Illinois Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）
Building Self-Evolving Agents via Experience-Driven Lifelong Learning: A Framework and Benchmark
School of Computer Science and Technology, East China Normal University（东华师范大学计算机科学与技术学院）；Shanghai AI Laboratory（上海人工智能实验室）；The Chinese University of HongKong（香港中文大学）；Fudan University（复旦大学）
DR Tulu: Reinforcement Learning with Evolving Rubrics for Deep Research
University of Washington（华盛顿大学）；Allen Institute for AI（人工智能研究院）；Carnegie Mellon University（卡内基梅隆大学）；Massachusetts Institute of Technology（麻省理工学院）；Seattle Children's Hospital（西雅图儿童医院）；University of California, Berkeley（加州大学伯克利分校）
Embodied Co-Design for Rapidly Evolving Agents: Taxonomy, Frontiers, and Challenges
Shenzhen International Graduate School (SIGS), Tsinghua University（深圳国际研究生院（SIGS），清华大学）；New Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences（模式识别新实验室（NLPR），自动化研究所，中国科学院）
Zhejiang University（浙江大学）；Huazhong University of Science and Technology（华中科技大学）；Xi’an Jiaotong University（西安交通大学）；National University of Defense Technology（国防科技大学）；Shanghai Jiao Tong University（上海交通大学）；Georgia Institute of Technology（佐治亚理工学院）；University of Science and Technology of China（中国科学技术大学）；Nanyang Technological University（南洋理工大学）；OPPO Research Institute（OPPO研究院）；Tianjin University（天津大学）；Chongqing University（重庆大学）；Alibaba Group（阿里巴巴集团）；Rakuten Group（乐天集团）
