Long-Horizon LLM Agents

LLM: Large Language Model

Overview

Long-horizon LLM agents are systems designed to pursue goals through many connected actions, often while using tools and responding to changing information. Common approaches include hierarchical planning and subgoals, reinforcement learning with step-level or milestone rewards, persistent memory and context compression, and harnesses that verify actions and recover from errors. Researchers evaluate them in interactive settings such as ALFWorld, WebShop, ScienceWorld, WebArena, software-engineering benchmarks, and deep-research tasks. A central challenge is making performance reliable over time: errors and lost constraints can compound, while memory, verification, and extra computation can improve success but add cost or introduce new failure modes.

Latest papers

CardsList

491 papers51 papers in the last four weeks, down 52% on the four weeks before. 0.4% of all new papers.

  1. Semantic Context Handoff with On-Demand Evidence for Long-Horizon Agents

    Nov 18, 2026Guolong Zheng, Chen Hou, Xu Yang +2Long-Horizon LLM AgentsLLM Context Management

  2. Communicative On-Policy Distillation: Distilling Black-Box Teacher Capabilities through Communication for Long-Horizon Tasks

    Nov 18, 2026Lirui Luo, Kelong Mao, Xinwei Yang +9On-Policy DistillationKnowledge Distillation

  3. From Repositories to Training Environments: Scaling Long-Horizon Coding Agents with Environment Evolution Loop

    Nov 16, 2026Yang Tian, Tianyu Zheng, Xiyuan Yang +5LLM Agent TrainingSoftware Engineering Agents

  4. CarryOn: Learning What to Keep in Context for Long-Horizon LLM Agents

    Nov 13, 2026Lihuan Li, Ruiyi Yang, Jiru Wang +5LLM Context ManagementLong-Horizon LLM Agents

  5. Folio: Recoverable Multi-Resolution Context Folding for Long-Horizon Language Agents

    Nov 12, 2026Yubo Ruan, Tao Peng, Jiao Zhang +11LLM Context ManagementContext Compaction for LLM Agents

  6. MemCredit: From Relevance to Utility for Decision-Time Memory in Long-Horizon Web Agents

    Nov 11, 2026Quanjiang Guo, Sijie Wang, Qianyang Zhu +4Memory Management for LLM AgentsLong-Horizon LLM Agents

  7. Incentive-Responsive Agents: A Principal-Agent Framework, Metric Suite, and Runtime Architecture for Long-Lived Personal AI Agents

    Nov 11, 2026Jipu SuLong-Horizon LLM AgentsLLM Agent Efficiency

  8. Keeping LLM Agents on Procedure: Executable Finite-State Control for Long-Horizon Instruction Fidelity

    Nov 7, 2026Yutao Sun, Haokun Li, Mingshuai Chen +4Long-Horizon LLM AgentsLLM Agent Reliability

  9. MemoryDojo: Benchmarking Poisoning Through LLM Agent Memory

    Nov 6, 2026David H. Yang, Blake Bullwinkel, Amanda Minnich +2Agent Memory PoisoningMemory Agent Benchmarks

  10. Rethinking Agent Trajectory Compression: A Workflow-Aware Perspective

    Nov 6, 2026Qijia Zhuang, Rui Liu, Wenqi Hu +1Context Compaction for LLM AgentsTask-Aware Compression

  11. SUMONE: A Simple and Efficient Summarizer for Plug-and-Play Agent Memory

    Nov 4, 2026Jizhe Zhang, Ang Li, Chenheng Zhang +1LLM Context CompressionContext Compaction for LLM Agents

  12. Adversarial Speculative Reasoning Improves Long Horizon Agents

    Nov 3, 2026Hongyin Luo, Wei Fang, Lijie Yang +5Speculative DecodingLong-Horizon LLM Reasoning

  13. Meta2^2: Precomputing Reusable Knowledge for Online Harness Evolution in Long-Horizon Agents

    Nov 3, 2026Lingzhi Yuan, Yiguang Cai, Weize Liu +5Agent Harness OptimizationAgent Harness Self-Evolution

  14. ToolSpace: Scaling Tool-Centric Environment Synthesis for Long-Horizon Agents

    Nov 3, 2026Guanting Dong, Zhichao Sun, Xinyu Yang +22LLM Tool UseLong-Horizon LLM Agents

  15. From Local Skills to Long-Horizon Web Tasks: Progressive Skill Exploration for LLM Web Agents

    Nov 2, 2026Xueyang Feng, Xiaohe Bo, Xu Chen +2LLM Agent Skill LearningLLM Agents

  16. BLPO: Beyond-Local Policy Optimization for Long-Horizon LLM Agents

    Nov 2, 2026Xuanren Chen, Chongyang Tao, Wenpeng Hu +4RL for LLM AgentsLong-Horizon LLM Agents