mlsanityData to 7 Oct 2026

Self-Improvements in Modern Agentic Systems: A Survey

arXiv 2607.13104 ↗First posted 14 Jul 2026cs.AI cs.CL cs.LGcited by 17

On the map: LLM agents and tool use → General LLM agent frameworks and benchmarks

also in 1 more
Abstract (arXiv)

Self-improving autonomous agents are moving from research prototypes to deployed systems. The primary goal is controllable evolution, or adaptation, from experience with minimal or even no human input. This survey frames modern self-improving agents as adaptive systems that convert experience into accumulated capability gains. We offer a system-level framework that represents a modern agent as a configuration coupling a foundation model with an operational scaffold of prompts, memory, tools, and control logic. Within this framework, self-improvement is formalized as a self-induced update operator that obtains and commits updates to model parameters or scaffold components. We organize prior work by update target and by the signals that drive change, then review applications and discuss evaluation, before closing with open problems and future directions. For convenience, we track technical updates on https://github.com/selfimproving-agent/awesome-Self-Improving-Agents.

Key facts. Cited by 17 papers in the map (data through 7 Oct 2026). Placed in General LLM agent frameworks and benchmarks, one of 16 directions in LLM agents and tool use. Builds on 171 of its 443 references that are in the map. First posted on arXiv 14 Jul 2026.

Builds on

171 of 443 references.

Closest ancestors · 20

Same direction, newest first

  1. 2026MetaGen: Self-Evolving Roles and Topologies for Multi-Agent LLM Reasoning27 Jan
  2. 2026Agent-as-a-Judge8 Jan
  3. 2025Alita-G: Self-Evolving Generative Agent for Agent Generation27 Oct
  4. 2025DeepAgent: A General Reasoning Agent with Scalable Toolsets24 Oct
  5. 2025Beyond Pipelines: A Survey of the Paradigm Shift toward Model-Native Agentic AI19 Oct
  6. 2025Self-Improving LLM Agents at Test-Time9 Oct
  7. 2025A Comprehensive Survey of Self-Evolving AI Agents: A New Paradigm Bridging Foundation Models and Lifelong Agentic Systems10 Aug
  8. 2025SEAgent: Self-Evolving Computer Use Agent with Autonomous Learning from Experience6 Aug
  9. 2025Evolving in Tasks: Empowering the Multi-modality Large Language Model as the Computer Use Agent6 Aug
  10. 2025MetaAgent: Toward Self-Evolving Agent via Tool Meta-Learning1 Aug
  11. 2025AgentDistill: Training-Free Agent Distillation with Generalizable MCP Boxes17 Jun
  12. 2025AgentOrchestra: Orchestrating Multi-Agent Intelligence with the Tool-Environment-Agent(TEA) Protocol14 Jun
8 more
  1. 2025TaskCraft: Automated Generation of Agentic Tasks11 Jun
  2. 2025Alita: Generalist Agent Enabling Scalable Agentic Reasoning with Minimal Predefinition and Maximal Self-Evolution26 May
  3. 2025VerifiAgent: a Unified Verification Agent in Language Model Reasoning1 Apr
  4. 2024How to Correctly do Semantic Backpropagation on Language-based Agentic Systems4 Dec
  5. 2024Agent-as-a-Judge: Evaluate Agents with Agents14 Oct
  6. 2024Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents13 Aug
  7. 2024AgentGym: Evolving Large Language Model-based Agents across Diverse Environments6 Jun
  8. 2023Conceptual Framework for Autonomous Cognitive Entities3 Oct

Other references · 151

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2021Evaluating Large Language Models Trained on Codecited by 8,053
  3. 2020Measuring Massive Multitask Language Understandingcited by 7,768
  4. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 2,950
  5. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,674
  6. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,081
  7. 2016Concrete Problems in AI Safetycited by 1,582
  8. 2018World Modelscited by 1,444
  9. 2023MemGPT: Towards LLMs as Operating Systemscited by 1,295
  10. 2019Dota 2 with Large Scale Deep Reinforcement Learningcited by 1,084
  11. 2024The AI Scientist: Towards Fully Automated Open-Ended Scientific Discoverycited by 1,002
  12. 2025A-MEM: Agentic Memory for LLM Agentscited by 1,000
38 more
  1. 2023DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelinescited by 803
  2. 2025AlphaEvolve: A coding agent for scientific and algorithmic discoverycited by 763
  3. 2017Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithmcited by 723
  4. 2025Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memorycited by 722
  5. 2024Self-Rewarding Language Modelscited by 654
  6. 2024InjecAgent: Benchmarking Indirect Prompt Injections in Tool-Integrated Large Language Model Agentscited by 476
  7. 2024Agentless: Demystifying LLM-based Software Engineering Agentscited by 400
  8. 2025The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Searchcited by 375
  9. 2025Zep: A Temporal Knowledge Graph Architecture for Agent Memorycited by 372
  10. 2025RAGEN: Understanding Self-Evolution in LLM Agents via Multi-Turn Reinforcement Learningcited by 357
  11. 2025Accelerating scientific discovery with Co-Scientistcited by 353
  12. 2023Check Your Facts and Try Again: Improving Large Language Models with External Knowledge and Automated Feedbackcited by 351
  13. 2025Agentic Context Engineering: Evolving Contexts for Self-Improving Language Modelscited by 349
  14. 2023The Curse of Recursion: Training on Generated Data Makes Models Forgetcited by 343
  15. 2025Absolute Zero: Reinforced Self-play Reasoning with Zero Datacited by 323
  16. 2025PaperBench: Evaluating AI's Ability to Replicate AI Researchcited by 286
  17. 2025SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?cited by 254
  18. 2025ReasoningBank: Scaling Agent Self-Evolving with Reasoning Memorycited by 253
  19. 2024TextGrad: Automatic "Differentiation" via Textcited by 250
  20. 2025Agentic Retrieval-Augmented Generation: A Survey on Agentic RAGcited by 242
  21. 2023MetaTool Benchmark for Large Language Models: Deciding Whether to Use Tools and Which to Usecited by 209
  22. 2024OS-Copilot: Towards Generalist Computer Agents with Self-Improvementcited by 203
  23. 2025Learning to Reason without External Rewardscited by 199
  24. 2025SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skillscited by 174
  25. 2025MIRIX: Multi-Agent Memory System for LLM-Based Agentscited by 160
  26. 2025Dynamic Cheatsheet: Test-Time Learning with Adaptive Memorycited by 146
  27. 2025MemOS: A Memory OS for AI Systemcited by 129
  28. 2026SkillOpt: Executive Strategy for Self-Evolving Agent Skillscited by 122
  29. 2025From Human Memory to AI Memory: A Survey on Memory Mechanisms in the Era of LLMscited by 119
  30. 2025Memento: Fine-tuning LLM Agents without Fine-tuning LLMscited by 119
  31. 2024GTBench: Uncovering the Strategic Reasoning Limitations of LLMs via Game-Theoretic Evaluationscited by 118
  32. 2026MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic Memorycited by 113
  33. 2023Mindstorms in Natural Language-Based Societies of Mindcited by 105
  34. 2024SafeAgentBench: A Benchmark for Safe Task Planning of Embodied LLM Agentscited by 102
  35. 2025MemGen: Weaving Generative Latent Memory for Self-Evolving Agentscited by 99
  36. 2023SCM: Enhancing Large Language Model with Self-Controlled Memory Frameworkcited by 98
  37. 2024SWE-Bench+: Enhanced Coding Benchmark for LLMscited by 95
  38. 2025Live-SWE-agent: Can Software Engineering Agents Self-Evolve on the Fly?cited by 91

50 most cited of 151 All 151 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 17.

17 papers.

Sort:

By area:

  1. 2026PAST-Bench: Benchmarking the Foundations of Recursive Self-Improvement in Personal Agentscited by 5
  2. 2026AutoDesign: Meta-Harness Optimization for Long-Horizon Agentic Designcited by 5
  3. 2026The Last AI Built by Humans: Toward Genuine Recursive Self-Improvementcited by 5
  4. 2026Auditing Harness Tampering in Self-Improving Agentscited by 3
  5. 2026Co-Evolution in Agentic Systems: Toward Self-Directed Evolution Beyond Human Designcited by 1
  6. 2026Agents in the Large: Perception-Centered Architecture for Persistent Agentscited by 1
  7. 2026MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselvescited by 1
  8. 2026NeoHorse-1: Towards Recursive Self-Improvement via Agentic Post-Training with Routing Harnesscited by 1
  9. 2026Efficiency Matters in Autonomous Researchcited by 0
  10. 2026Reflections on Trusting Trust, Revisited: Contaminating Self-Modifying AI Coding Agents with Poisoned Benchmarkscited by 0
  11. 2026Evolutionary Safety of Recursive Self-Improving AI: Taxonomy, Risk Discovery, and Evaluationcited by 0
  12. 2026PluginRSI: Recursive Improvement of Agent Harnesses with Reusable Pluginscited by 0
5 more
  1. 2026R2 Flow: Recursive Self-Improvement via Recursive Skill Evolutioncited by 0
  2. 2026Self-Evolving Agents via Likelihood-Guided Tool-Space Optimizationcited by 0
  3. 2026AutoDataBench: Can Agents Write the Data That Feeds the Self-Improvement Loop?cited by 0
  4. 2026Mimir: Physics-Grounded LLM Agents for Long-Horizon Irrigation Controlcited by 0
  5. 2026ReSI: Recursive Safety Improvement toward Resistant and Resilient AIcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Self-Aware Recursively Self-Improving Agents for Personal Singularity: A Goal-, Scope-, Tool-, and Benchmark-Driven Multi-Agent Architecture
  2. 2026Hierarchical Self-Improvement: A Framework for Task-Specific Evolvable Agent Harnesses
  3. 2026MOSS: Self-Evolution through Source-Level Rewriting in Autonomous Agent Systems
  4. 2026Agent Plasticity: Measuring Self-Improvement Through Experience
  5. 2025A Survey of Self-Evolving Agents: What, When, How, and Where to Evolve on the Path to Artificial Super Intelligence
  6. 2026Rethinking Self-Evolving Agent Skills: Feedback Dynamics over Multiple Rounds
  7. 2026Knowledge-Centric Self-Improvement
  8. 2026RRSI: Regularized Recursive Self-Improvement of Agent Harnesses