A Survey of Self-Evolving Agents: On Path to Artificial Super Intelligence

  • 2025-07-30 17:59:37
  • Huan-ang Gao, Jiayi Geng, Wenyue Hua, Mengkang Hu, Xinzhe Juan, Hongzhang Liu, Shilong Liu, Jiahao Qiu, Xuan Qi, Yiran Wu, Hongru Wang, Han Xiao, Yuhang Zhou, Shaokun Zhang, Jiayi Zhang, Jinyu Xiang, Yixiong Fang, Qiwen Zhao, Dongrui Liu, Qihan Ren, Cheng Qian, Zhenghailong Wang, Minda Hu, Huazheng Wang, Qingyun Wu, Heng Ji, Mengdi Wang
  • 0

Abstract

Large Language Models (LLMs) have demonstrated strong capabilities but remainfundamentally static, unable to adapt their internal parameters to novel tasks,evolving knowledge domains, or dynamic interaction contexts. As LLMs areincreasingly deployed in open-ended, interactive environments, this staticnature has become a critical bottleneck, necessitating agents that canadaptively reason, act, and evolve in real time. This paradigm shift -- fromscaling static models to developing self-evolving agents -- has sparked growinginterest in architectures and methods enabling continual learning andadaptation from data, interactions, and experiences. This survey provides thefirst systematic and comprehensive review of self-evolving agents, organizedaround three foundational dimensions -- what to evolve, when to evolve, and howto evolve. We examine evolutionary mechanisms across agent components (e.g.,models, memory, tools, architecture), categorize adaptation methods by stages(e.g., intra-test-time, inter-test-time), and analyze the algorithmic andarchitectural designs that guide evolutionary adaptation (e.g., scalar rewards,textual feedback, single-agent and multi-agent systems). Additionally, weanalyze evaluation metrics and benchmarks tailored for self-evolving agents,highlight applications in domains such as coding, education, and healthcare,and identify critical challenges and research directions in safety,scalability, and co-evolutionary dynamics. By providing a structured frameworkfor understanding and designing self-evolving agents, this survey establishes aroadmap for advancing adaptive agentic systems in both research and real-worlddeployments, ultimately shedding lights to pave the way for the realization ofArtificial Super Intelligence (ASI), where agents evolve autonomously,performing at or beyond human-level intelligence across a wide array of tasks.

 

Quick Read (beta)

loading the full paper ...