Date
Publisher
arXiv
Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse tasks but remain fundamentally static, unable to adapt their internal parameters to novel tasks, evolving knowledge domains, or dynamic interaction contexts. As LLMs are increasingly deployed in open-ended, interactive environments, this static nature has become a critical bottleneck, necessitating agents that can adaptively reason, act, and evolve in real time. This paradigm shift -- from scaling static models to developing self-evolving agents -- has sparked growing interest in architectures and methods enabling continual learning and adaptation from data, interactions, and experiences. This survey provides the first systematic and comprehensive review of self-evolving agents, organizing the field around three foundational dimensions: what, when, and how to 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 and architectural designs that guide evolutionary adaptation (e.g., scalar rewards, textual feedback, single-agent and multi-agent systems). Additionally, we analyze 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 framework for understanding and designing self-evolving agents, this survey establishes a roadmap for advancing more adaptive, robust, and versatile agentic systems in both research and real-world deployments, and ultimately sheds light on the realization of Artificial Super Intelligence (ASI) where agents evolve autonomously and perform beyond human-level intelligence across tasks.
Study at a glance
- Question: How can artificial intelligence agents adapt, learn, and evolve continuously over time?
- Setting and data: Literature review of artificial intelligence research papers and benchmarks published between 2022 and 2025.
- Method: Systematic literature review and taxonomy synthesis (Systematic / scoping review).
Why it matters
Static AI models cannot adapt their parameters to novel tasks or dynamic contexts. Developing self-evolving agents capable of continuous adaptation from real-world interactions addresses this critical bottleneck.
Key findings
- Agent evolution operates across four components—models, context, tools, and architecture—occurring either during task execution (intra-test-time) or between completed tasks (inter-test-time).
- Self-evolution methods rely primarily on reward-based signals, imitation learning from generated demonstrations, or population-based evolutionary search and self-play.
- Current evaluation practices largely rely on static assessments, revealing major gaps in measuring long-term knowledge retention, catastrophic forgetting, operational costs, and safety drift.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
What is the application?
Who age?
Study design
