Self-Evolving Agents

Follow this topic Ask about this topic

Self-evolving agents are systems that use experience from task execution to change how they solve future tasks, rather than relying only on a fixed model or hand-designed workflow. Common approaches store and refine reusable skills or memories, optimize prompts and agent harnesses, train policies with reinforcement learning, or co-evolve task generators, environments, and teams of agents. These methods are tested in settings such as web and GUI use, coding, embodied robotics, scientific workflows, and reasoning benchmarks, with feedback ranging from execution outcomes to verifiers and paired validation. A central challenge is making improvements transfer to new tasks without forgetting earlier abilities or introducing unsafe behavior; reliable evaluation, update verification, and long-term monitoring remain open problems.

Latest papers

427 papers104 papers in the last four weeks, up 148% on the four weeks before. 0.6% of all new papers.

StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability?
Sep 1, 2026Yinghao Chen, Zixi Chen, Bingxiang He +7Benchmark Design Language Model Self-Improvement
Tsinghua University · Zhejiang University
Aspire: Can Models Self-Evolve from Vague Goals?
Aug 31, 2026Yuhao Wu, Jingyuan Zhang, Jiajun Shi +18LLM Agent Self-Improvement AI Agent Benchmarks
PRACTICE: From Experience to Expertise in Self-Evolving Embodied Agents
Aug 31, 2026Ziyi Bai, Siqi Li, Tinglei Huang +1Continual Learning for LLM Agents Robot Skill Learning
Beijing Academy of Artificial Intelligence (BAAI) · Institute of Software, Chinese Academy of Sciences · University of Chinese Academy of Sciences
EvoSkill Injection: Red-Teaming Autonomous Skill Generation and Evolution in Self-Evolving Agents
Aug 31, 2026Doyun Kim, Chanwoo Kim, Sugyeong Eo +2AI Agent Security AI Agent Security Benchmarks
Soongsil University · Yonsei University · Jeju National University
Beyond Uncertainty: Multi-Solver Disagreement Rewards for Self-Evolving Reasoning Curricula
Aug 30, 2026Vinoth Selvendran, Zhanming ZhangReward Modeling RL for Language Model Reasoning
Independent Researcher · Palo Alto, CA, USA · New York, NY, USA
EvoUndo: Recoverability-Constrained Self-Evolution for LLM Agent Harnesses
Aug 28, 2026Tanmay Sah, Dolly Sah, Harshul Jain +1LLM Agent Self-Improvement LLM Agent Harnesses
Independent Researcher
RedEvoAgent: Automatic Red-Teaming Agent with Experience-Driven Skill Evolution
Aug 27, 2026Junjie Zhang, Hui Liu, Kecheng Chen +3Automated Red Teaming LLM Agent Security
City University of Hong Kong · Shenzhen MSU-BIT University
SwarmWorld: Stigmergic technological evolution in societies of language-model agents
Aug 26, 2026Subhadeep Pal, Fiona Y. Wang, Markus J. BuehlerEmergent Behavior in Multi-Agent Systems Multi-Agent Coordination
Laboratory for Atomistic and Molecular Mechanics (LAMM), Massachusetts Institute of Technology, Cambridge, MA 02139, USA · Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA · Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA +2
EVOMAL: Self-Poisoning in Self-Evolving Coding Agents
Aug 26, 2026Xiaodong Wu, Yu Shi, Qi Li +5LLM Agent Security Coding Agents
Queen’s University
From General Agents to RCA Experts: A Self-Evolving Harness for Root Cause Analysis
Aug 26, 2026Haiyu Huang, Jiewei Lyu, Zhihan Jiang +5Root Cause Analysis LLM Agent Harnesses
The Chinese University of Hong Kong · Individual Researcher · ByteDance
StarHarness: Evolving Harnesses with Stratified Search for Enterprise Environments
Aug 25, 2026Esakkivel Esakkiraja, Denis Akhiyarov, Vikas Yadav +4Agent Harness Optimization LLM Agent Harnesses
ServiceNow · Mila · Université de Montréal
MediSkill-Evo: Process-Constrained Self-Evolution for Evidence-Grounded Clinical Interaction
Aug 24, 2026Ruoyu Wu, Shenfu Xie, Yinqian Sun +2Self-Evolving Agents LLM Agent Skill Learning
Brain-inspired Cognitive AI Lab, Institute of Automation, Chinese Academy of Sciences · Beijing Key Laboratory of Safe AI and Superalignment · Beijing Institute of AI Safety and Governance +2
Coalition-Aware Skill Reliability for Self-Evolving Agents
Aug 23, 2026Qiyan Zhao, Xiaofeng Zhang, Bo Liu +11LLM Agent Skill Learning LLM Agent Reliability
CASIA · LongShine AI Lab · SJTU +2
CONTRAMEM: Learning Self-Evolving Procedural Memory from Contrasting Multi-Model Trajectories
Aug 23, 2026Zheyuan Deng, Binghang Lu, Hanqi Feng +11Computer-Use Agents LLM Agent Memory
Brown University · Purdue University · Carnegie Mellon University +6
SSE-Bio: A Structured Self-Evolving Agent with Agentic Retrieval Policy for Multi-Hop Biomedical Reasoning
Aug 22, 2026Zhaohan Meng, Zaiqiao Meng, Siwei Liu +3Biomedical QA Self-Evolving Agents
School of Computing Science, University of Glasgow · Brigham and Women’s Hospital, Harvard Medical School, Harvard University · Language Technology Lab, University of Cambridge +3
OptiMAS: Automatically Optimize Multi-Agent System
Aug 22, 2026Yuxin Cheng, Chang Liu, Hanxin Yu +9Multi-Agent System Optimization Multi-Agent LLM Systems
The University of Hong Kong · Huawei Foundation Model Department, Huawei Technologies Ltd. · City University of Hong Kong +2
ForeDreamer: A Self-Evolving Dual-Agent Memory Architecture for Future Event Prediction
Aug 21, 2026Linhao Zhong, Zongze Du, Linyu Wu +6Memory-Augmented Agents Self-Evolving Agents
Zhejiang University, State Lab of CAD & CG · Ant Group · National University of Singapore +1
EvoTS-Agent: A Self-Evolving LLM Agent for Financial Time Series Change Point Detection
Aug 18, 2026Lei Jiang, Ye Wei, Xinyu Xi +7Change-Point Detection Self-Evolving Agents
Alan Turing Institute · University of Oxford · National University of Singapore +3
Auditing Self-Evolution in Financial Agents: Capability Gains, Security Drift, and Execution-Interface Mismatch
Aug 18, 2026Jialong Li, Jialing ZhuAI Agent Auditing AI Agent Security
Independent Researcher
Zetta $ζ$ : An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence
Aug 17, 2026Xin Ding, Liang Mi, Mingzhe Huang +12Embodied AI Self-Evolving Agents
Institute for AI Industry Research (AIR), Tsinghua University; Z-Trans AI
VCE-Skill: Enhancing Skill Self-Evolution with Version-Change Experience
Aug 17, 2026Jianming Chen, Xuanbin Ye, Yawen Wang +3LLM Agent Skill Learning LLM Agent Self-Improvement
Institute of Software Chinese Academy of Sciences, Beijing, China · Science & Technology on Integrated Information System Laboratory, Beijing, China · State Key Laboratory of Complex System Modeling and Simulation Technology, Beijing, China +2
Competing at Every Price Point with Agentic Evolution over a Menu of LLMs
Aug 17, 2026Andrew BorthwickLLM Agents LLM Agent Training
Independent Researcher
Evo-Harness: Context-to-Harness Skill Compilation for Self-Evolving Agents
Aug 15, 2026Tianxin Wei, Zhan Shi, Minhua Lin +14Continual Learning for LLM Agents LLM Agent Skill Learning
University of Illinois Urbana-Champaign · Amazon · The Pennsylvania State University +3
SysEvolve: An AI-native, safe, autonomous adversarial attack-defense co-evolutionary system
Aug 15, 2026Yuhan Meng, Shaofei Li, Jionghao Huang +8Cybersecurity Autonomous Cyber Defense
School of Computer Science, Peking University · Southeast University · School of Computing, National University of Singapore
PACE-Bench: Benchmarking Physics Adaptation via Code Evolution in Dynamic Environments
Aug 14, 2026Yuhao Zhan, Bingxiang He, Zecong Tang +1AI Agent Benchmarks Self-Evolving Agents
Tsinghua University · Zhejiang University
SkillEvo: Self-Renewing Evolution Gradients from Multi-Turn Interaction Feedback
Aug 13, 2026Qianxi Yan, Chunrong Chen, Jiuzhou Zhao +3LLM Agent Skill Learning Multi-Turn Dialogue Evaluation
†Tencent Cloud Andon · ‡Zhejiang University
Latent On-Policy Self-Distillation
Aug 13, 2026Guibin Zhang, Jiayang Lyu, Ran Sun +4On-Policy Self-Distillation Knowledge Distillation
1National University of Singapore · 2Beijing University of Posts and Telecommunications · 3Shanghai Jiao Tong University
Beyond Handcrafted Security: Towards Self-Evolving Defense for LLM Agents
Aug 13, 2026Jiajun Ruan, Peiyang Li, Yukun Chen +2LLM Agent Security LLM Agent Safety
University of Minnesota · Ant Group · Tsinghua University +1
Spatial Memory Agent: Experience-Grounded Procedural Memory for Spatial Intelligence
Aug 13, 2026Haokai Zhang, Yuhang Ding, Yunshu Zhou +5Memory-Augmented VLMs Visual Spatial Reasoning
Zhejiang University · Shanghai Jiao Tong University · Shanghai Innovation Institute
MindMemOS: A Portable and Self-Evolving Memory Operating Layer for AI Agents
Aug 12, 2026Kaichao Liang, Yuqi Cui, Hao Kong +13Agent Memory Agent Skill Learning
Noah’s Ark Lab, Huawei Technologies

Sep 1, 2026J/K move · Enter open · S save

StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability?

Yinghao Chen, Zixi Chen, Bingxiang He, Ziqing Qiao+6

Tsinghua University · Zhejiang University

Benchmark Design Language Model Self-Improvement Self-Evolving Agents

Open paper AskPDF arXiv

Humans need to study only a handful of well-written textbooks to master a discipline and attempt its hardest problems. We argue that an ideal self-evolution method should share the same property, that is autonomously learning from raw training material for transferable problem-solving capability. However, we still lack a direct measurement for it. We introduce StudyBench, a controlled physics benchmark that directly measures how efficiently a self-evolution method converts training material into capability. We organise the test set into an Application Set, consisting of difficult textbook problems and evaluating absorption ability, and a Transfer Set, consisting of olympiad-level problems and evaluating transfer ability. Benchmarking representative self-evolution methods across three base models, we find that improvements on the Application Set rarely translate to the harder Transfer Set. A guidance ablation exposes a Guidance Gap: even the strongest method closes only a small fraction of what the same material unlocks when supplied as in-context guidance. Besides, every method hits a Compute Plateau, saturating well before exhausting its compute budget. The remaining gap is therefore a method problem rather than a data or compute problem. By offering a clean and controlled benchmark, StudyBench turns self-evolution progress from an open-ended pursuit into a measurable target for future research. Our code is released at https://github.com/thunlp/StudyBench.