KLong通过分段微调与渐进强化学习，让大模型搞定超长任务。
KLong: Training LLM Agent for Extremely Long-horizon Tasks
- 用分段微调保留早期上下文，逐步截断后期内容并保持重叠。
- 在PaperBench上比Kimi K2 Thinking（1T）高11.28%，跨任务泛化能力强。
- 适合需要长期规划与复杂推理的智能体应用，如科研辅助与自动化编程。
本文提出KLong，一个开源的大语言模型智能体，专为解决超长时序任务而设计。核心方法是先通过轨迹分段监督微调（SFT）冷启动模型，再通过渐进式强化学习（RL）扩展能力。首先，采用全面的SFT方案激活基础模型的智能体能力；随后引入Research-Factory自动化流水线，通过收集研究论文并构建评估标准生成高质量训练数据。基于Claude 4.5 Sonnet（Thinking）生成数千条长轨迹，并设计新的轨迹分段SFT方法，该方法保留早期上下文、逐步截断后期内容并维持子轨迹间的重叠。为进一步提升长时序任务解决能力，提出一种新型渐进式强化学习，将训练分为多个阶段，逐步延长超时时间。实验表明，KLong（106B）在PaperBench上超越Kimi K2 Thinking（1T）达11.28%，性能提升在SWE-bench Verified和MLE-bench等编码基准上同样显著。
原文摘要 · Abstract (English)
This paper introduces KLong, an open-source LLM agent trained to solve extremely long-horizon tasks. The principle is to first cold-start the model via trajectory-splitting SFT, then scale it via progressive RL training. Specifically, we first activate basic agentic abilities of a base model with a comprehensive SFT recipe. Then, we introduce Research-Factory, an automated pipeline that generates high-quality training data by collecting research papers and constructing evaluation rubrics. Using this pipeline, we build thousands of long-horizon trajectories distilled from Claude 4.5 Sonnet (Thinking). To train with these extremely long trajectories, we propose a new trajectory-splitting SFT, which preserves early context, progressively truncates later context, and maintains overlap between sub-trajectories. In addition, to further improve long-horizon task-solving capability, we propose a novel progressive RL, which schedules training into multiple stages with progressively extended timeouts. Experiments demonstrate the superiority and generalization of KLong, as shown in Figure 1. Notably, our proposed KLong (106B) surpasses Kimi K2 Thinking (1T) by 11.28% on PaperBench, and the performance improvement generalizes to other coding benchmarks like SWE-bench Verified and MLE-bench.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。