Elamur Research⌕
☰
home / papers / 2026 / MemoBrain: Executive Memory as an Agentic Brain for Reasoning

MemoBrain: Executive Memory as an Agentic Brain for Reasoning

Paper · 2026 · Hongjin Qian, Zhao Cao and Zheng Liu

Summary

Tool-using LLM agents on long-horizon tasks accumulate reasoning traces and tool outputs. These fill their limited context window with noise, which weakens coherence and goal alignment. The authors propose "executive memory" as a remedy. It is a per-task memory that starts empty, evolves online with the reasoning, and actively controls what the agent sees. Their system, MemoBrain, runs as a separate copilot model beside the reasoning agent. In the background, it condenses each finished episode into a "thought" that records the subproblem, the tools or evidence used, and the outcome. Explicit dependency links connect each thought to earlier ones, forming a reasoning graph. When a fixed context budget is reached, MemoBrain folds completed sub-trajectories and flushes invalid or low-value steps, keeping only a compact reasoning backbone. Memory construction and memory management each receive task-specific training. MemoBrain was attached to strong agents, with Qwen3-based memory models from 4B to 14B. It consistently outperformed baselines on GAIA, WebWalker and BrowseComp-Plus. Its significance is that it treats memory as a reusable control module rather than mere context compression.

Key points

  • Three memory operations: consulting the global memory, sequential folding of finished episodes, and selective memory flush. Memory units are typically about 10× smaller than the raw traces they replace.
  • Memory backbones tested were Qwen3 4B, 8B and 14B. DeepSeek-V3.2 as the memory model gave the best results, but the 8B and 14B versions came close, so MemoBrain-8B is the default. Untuned Qwen3-14B produced invalid memory updates, suggesting targeted training matters more than model size.
  • Budget sensitivity: budgets under 16K can fall below the no-memory baseline, and gains level off above 64K.
  • Efficiency was measured on BrowseComp-Plus with a 32K memory budget, using 4×H100 for the reasoner and 2×H100 for the memory model. Asynchronous memorization adds negligible latency up to 128K tokens and stays non-blocking beyond 256K.
  • The base agent is capped at a 128K window. With MemoBrain, it reasons over hundreds of thousands of tokens and makes many more tool calls.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Renmin University of ChinaBeijing · China114 works · 16,118 citations🎓Hong Kong Polytechnic UniversityHong Kong · China84 works · 4,956 citations🔬Beijing Academy of Artificial IntelligenceBeijing · China32 works · 7,511 citations

 Fields, topics and concepts

📚 Shelves 3The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksInformation Systems · 1,437 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,426 works

Cite

Cite as Hongjin Qian, Zhao Cao and Zheng Liu. 2026. MemoBrain: Executive Memory as an Agentic Brain for Reasoning. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.127

@inproceedings{qian2026memobrain,
  title = {{MemoBrain: Executive Memory as an Agentic Brain for Reasoning}},
  author = {Hongjin Qian, Zhao Cao and Zheng Liu},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.127},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.127}
}

References

49 references, 20 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Yushi Bai, Shangqing Tu, Jiajie Zhang, Hao Peng, Xiaozhi Wang, Xin Lv, Shulin Cao, Jiazheng Xu, Lei Hou, Yuxiao Dong, Jie Tang, and Juanzi Li. 2025. Longbench v2: Towards deeper understanding and reasoning on realistic long-context multitasks. Preprint, arXiv:2412.15204.
  2. Chi-Min Chan, Chunpu Xu, Ruibin Yuan, Hongyin Luo, Wei Xue, Yike Guo, and Jie Fu. 2024. Rq-rag: Learning to refine queries for retrieval augmented generation. In First Conference on Language Modeling. find
  3. Zijian Chen, Xueguang Ma, Shengyao Zhuang, Ping Nie, Kai Zou, Andrew Liu, Joshua Green, Kshama Patel, Ruoxi Meng, Mingyi Su, Sahel Sharifymoghaddam, Yanxi Li, Haoran Hong, Xinyu Shi, Xuye Liu, Nandan Thakur, Crystina Zhang, Luyu Gao, Wenhu Chen, and Jimmy Lin. 2025. Browsecomp-plus: A more fair and… — [arXiv:2508.06600](https://arxiv.org/abs/2508.06600) find
  4. Prateek Chhikara, Dev Khant, Saket Aryan, Taranjeet Singh, and Deshraj Yadav. 2025. Mem0: Building production-ready AI agents with scalable long-term memory. CoRR, abs/2504.19413.
  5. DeepSeek-AI. 2025. Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. CoRR, abs/2501.12948.
  6. Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, and Jonathan Larson. 2024. From local to global: A graph rag approach to query-focused summarization. Preprint, arXiv:2404.16130. find
  7. 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, Zhenhailong Wang, Minda Hu, Huazheng W… find
  8. Team GLM, Aohan Zeng, Bin Xu, Bowen Wang, Chenhui Zhang, Da Yin, Diego Rojas, Guanyu Feng, Hanlin Zhao, Hanyu Lai, Hao Yu, Hongning Wang, Jiadai Sun, Jiajie Zhang, Jiale Cheng, Jiayi Gui, Jie Tang, Jing Zhang, Juanzi Li, Lei Zhao, Lindong Wu, Lucen Zhong, Mingdao Liu, Minlie Huang, Peng Zhang, Qinka… — ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools
  9. Cheng-Ping Hsieh, Simeng Sun, Samuel Kriman, Shantanu Acharya, Dima Rekesh, Fei Jia, and Boris Ginsburg. 2024. Ruler: What’s the real context size of your long-context language models? Preprint, arXiv:2404.06654.
  10. Yuyang Hu, Shichun Liu, Yanwei Yue, Guibin Zhang, Boyang Liu, Fangyi Zhu, Jiahang Lin, Honglin Guo, Shihan Dou, Zhiheng Xi, Senjie Jin, Jiejun Tan, Yanbin Yin, Jiongnan Liu, Zeyu Zhang, Zhongxiang Sun, Yutao Zhu, Hao Sun, Boci Peng, Zhenrong Cheng, Xuanbo Fan, Jiaxin Guo, Xinlei Yu, Zhenhong Zhou, Z… — [arXiv:2512.13564](https://arxiv.org/abs/2512.13564) find
  11. Jiazheng Kang, Mingming Ji, Zhe Zhao, and Ting Bai. 2025. Memory os of ai agent. Preprint, arXiv:2506.06326.
  12. Baixuan Li, Bo Zhang, Dingchu Zhang, Fei Huang, Guangyu Li, Guoxin Chen, Huifeng Yin, Jialong Wu, Jingren Zhou, Kuan Li, Liangcai Su, Litu Ou, Liwen Zhang, Pengjun Xie, Rui Ye, Wenbiao Yin, Xinmiao Yu, Xinyu Wang, Xixi Wu, Xuanzhong Chen, Yida Zhao, Zhen Zhang, Zhengwei Tao, Zhongwang Zhang, Zile Qi… — [arXiv:2510.24701](https://arxiv.org/abs/2510.24701) find
  13. Xiaoxi Li, Guanting Dong, Jiajie Jin, Yuyao Zhang, Yujia Zhou, Yutao Zhu, Peitian Zhang, and Zhicheng Dou. 2025b. Search-o1: Agentic search-enhanced large reasoning models. CoRR, abs/2501.05366.
  14. Xiaoxi Li, Wenxiang Jiao, Jiarui Jin, Guanting Dong, Jiajie Jin, Yinuo Wang, Hao Wang, Yutao Zhu, Ji-Rong Wen, Yuan Lu, and Zhicheng Dou. 2025c. Deepagent: A general reasoning agent with scalable toolsets. CoRR, abs/2510.21618. — [arXiv:2510.21618](https://arxiv.org/abs/2510.21618) find
  15. Xiaoxi Li, Jiajie Jin, Guanting Dong, Hongjin Qian, Yutao Zhu, Yongkang Wu, Ji-Rong Wen, and Zhicheng Dou. 2025d. Webthinker: Empowering large reasoning models with deep research capability. arXiv preprint arXiv:2504.21776. find
  16. Zhiyu Li, Shichao Song, Hanyu Wang, Simin Niu, Ding Chen, Jiawei Yang, Chenyang Xi, Huayi Lai, Jihao Zhao, Yezhaohui Wang, Junpeng Ren, Zehao Lin, Jiahao Huo, Tianyi Chen, Kai Chen, Kehang Li, Zhiqiang Yin, Qingchen Yu, Bo Tang, Hongkang Yang, Zhi-Qin John Xu, and Feiyu Xiong. 2025e. Memos: An opera… — [arXiv:2505.22101](https://arxiv.org/abs/2505.22101) find
  17. Adyasha Maharana, Dong-Ho Lee, Sergey Tulyakov, Mohit Bansal, Francesco Barbieri, and Yuwei Fang. 2024. Evaluating very long-term conversational memory of llm agents. Preprint, arXiv:2402.17753. find
  18. Lingrui Mei, Jiayu Yao, Yuyao Ge, Yiwei Wang, Baolong Bi, Yujun Cai, Jiazhi Liu, Mingyu Li, Zhong-Zhi Li, Duzhen Zhang, et al. 2025. A survey of context engineering for large language models. arXiv preprint arXiv:2507.13334. — [arXiv:2507.13334](https://arxiv.org/abs/2507.13334) find
  19. Grégoire Mialon, Clémentine Fourrier, Thomas Wolf, Yann LeCun, and Thomas Scialom. 2023. Gaia: a benchmark for general ai assistants. In The Twelfth International Conference on Learning Representations.
  20. OpenAI. 2024. Gpt-4 technical report. arXiv:2303.08774.
  21. Preprint, Charles Packer, Vivian Fang, Shishir G Patil, Kevin Lin, Sarah Wooders, and Joseph E Gonzalez. 2023. Memgpt: Towards llms as operating systems. arXiv preprint arXiv:2310.08560. find
  22. Cheng Qian, Emre Can Acikgoz, Qi He, Hongru Wang, Xiusi Chen, Dilek Hakkani-Tür, Gokhan Tur, and Heng Ji. 2025a. Toolrl: Reward is all tool learning needs. Preprint, arXiv:2504.13958. find
  23. Hongjin Qian and Zheng Liu. 2025a. Metaagent: Toward self-evolving agent via tool meta-learning. Preprint, arXiv:2508.00271. — [arXiv:2508.00271](https://arxiv.org/abs/2508.00271) find
  24. Hongjin Qian and Zheng Liu. 2025b. Scent of knowledge: Optimizing search-enhanced reasoning with information foraging. arXiv preprint arXiv:2505.09316. — [arXiv:2505.09316](https://arxiv.org/abs/2505.09316) find
  25. Hongjin Qian, Zheng Liu, Peitian Zhang, Zhicheng Dou, and Defu Lian. 2025b. Boosting long-context information seeking via query-guided activation refilling. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 9453–9464. find
  26. Hongjin Qian, Zheng Liu, Peitian Zhang, Kelong Mao, Defu Lian, Zhicheng Dou, and Tiejun Huang. 2025c. Memorag: Boosting long context processing with global memory-enhanced retrieval augmentation. In Proceedings of the ACM on Web Conference 2025, WWW ’25, page 2366–2377, New York, NY, USA. Associatio… find
  27. Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D Manning, Stefano Ermon, and Chelsea Finn. 2024. Direct preference optimization: Your language model is secretly a reward model. Advances in Neural Information Processing Systems, 36.
  28. Zhihong Shao, Yeyun Gong, Yelong Shen, Minlie Huang, Nan Duan, and Weizhu Chen. 2023. Enhancing retrieval-augmented large language models with iterative retrieval-generation synergy. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 9248–9274.
  29. Weiwei Sun, Miao Lu, Zhan Ling, Kang Liu, Xuesong Yao, Yiming Yang, and Jiecao Chen. 2025. Scaling long-horizon llm agent via context-folding. Preprint, arXiv:2510.11967. — [arXiv:2510.11967](https://arxiv.org/abs/2510.11967) find
  30. Lei Wang, Chen Ma, Xueyang Feng, Zeyu Zhang, Hao Yang, Jingsen Zhang, Zhiyuan Chen, Jiakai Tang, Xu Chen, Yankai Lin, Wayne Xin Zhao, Zhewei Wei, and Jirong Wen. 2024. A survey on large language model based autonomous agents. Frontiers Comput. Sci., 18(6):186345.
  31. Weizhi Wang, Li Dong, Hao Cheng, Xiaodong Liu, Xifeng Yan, Jianfeng Gao, and Furu Wei. 2023. Augmenting language models with long-term memory. In Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Systems 2023, NeurIPS 2023, New Orleans, LA, USA,… find
  32. Jason Wei, Zhiqing Sun, Spencer Papay, Scott McKinney, Jeffrey Han, Isa Fulford, Hyung Won Chung, Alex Tachard Passos, William Fedus, and Amelia Glaese. 2025. Browsecomp: A simple yet challenging benchmark for browsing agents. arXiv preprint arXiv:2504.12516. find
  33. Jialong Wu, Baixuan Li, Runnan Fang, Wenbiao Yin, Liwen Zhang, Zhengwei Tao, Dingchu Zhang, Zekun Xi, Yong Jiang, Pengjun Xie, Fei Huang, and Jingren Zhou. 2025a. Webdancer: Towards autonomous information seeking agency. CoRR, abs/2505.22648.
  34. Jialong Wu, Wenbiao Yin, Yong Jiang, Zhenglin Wang, Zekun Xi, Runnan Fang, Linhai Zhang, Yulan He, Deyu Zhou, Pengjun Xie, et al. 2025b. Webwalker: Benchmarking llms in web traversal. arXiv preprint arXiv:2501.07572.
  35. Xixi Wu, Kuan Li, Yida Zhao, Liwen Zhang, Litu Ou, Huifeng Yin, Zhongwang Zhang, Yong Jiang, Pengjun Xie, Fei Huang, Minhao Cheng, Shuai Wang, Hong Cheng, and Jingren Zhou. 2025c. Resum: Unlocking long-horizon search intelligence via context summarization. CoRR, abs/2509.13313.
  36. Ziyi Xia, Kun Luo, Hongjin Qian, and Zheng Liu. 2025a. Open data synthesis for deep research. arXiv preprint arXiv:2509.00375. — [arXiv:2509.00375](https://arxiv.org/abs/2509.00375) find
  37. Ziyi Xia, Kun Luo, Hongjin Qian, and Zheng Liu. 2025b. Open data synthesis for deep research. Preprint, arXiv:2509.00375. — [arXiv:2509.00375](https://arxiv.org/abs/2509.00375) find
  38. Wujiang Xu, Zujie Liang, Kai Mei, Hang Gao, Juntao Tan, and Yongfeng Zhang. 2025. A-mem: Agentic memory for llm agents. Preprint, arXiv:2502.12110.
  39. B. Y. Yan, Chaofan Li, Hongjin Qian, Shuqi Lu, and Zheng Liu. 2025a. General agentic memory via deep research. Preprint, arXiv:2511.18423. — [arXiv:2511.18423](https://arxiv.org/abs/2511.18423) find
  40. Sikuan Yan, Xiufeng Yang, Zuchao Huang, Ercong Nie, Zifeng Ding, Zonggen Li, Xiaowen Ma, Kristian Kersting, Jeff Z. Pan, Hinrich Schütze, Volker Tresp, and Yunpu Ma. 2025b. Memory-r1: Enhancing large language model agents to manage and utilize memories via reinforcement learning. Preprint, arXiv:250…
  41. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, Chujie Zheng, Dayiheng Liu, Fan Zhou, Fei Huang, Feng Hu, Hao Ge, Haoran Wei, Huan Lin, Jialong Tang, Jian Yang, Jianhong Tu, Jianwei Zhang, Jian Yang, Jiaxi Yang, Jingren Zhou, Jun… — Qwen3 Technical Report
  42. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. 2023. React: Synergizing reasoning and acting in language models. In International Conference on Learning Representations (ICLR).
  43. Rui Ye, Zhongwang Zhang, Kuan Li, Huifeng Yin, Zhengwei Tao, Yida Zhao, Liangcai Su, Liwen Zhang, Zile Qiao, Xinyu Wang, Pengjun Xie, Fei Huang, Siheng Chen, Jingren Zhou, and Yong Jiang. 2025. Agentfold: Long-horizon web agents with proactive context management. CoRR, abs/2510.24699. — [arXiv:2510.24699](https://arxiv.org/abs/2510.24699) find
  44. Hongli Yu, Tinghong Chen, Jiangtao Feng, Jiangjie Chen, Weinan Dai, Qiying Yu, Ya-Qin Zhang, Wei-Ying Ma, Jingjing Liu, Mingxuan Wang, and Hao Zhou. 2025. Memagent: Reshaping long-context llm with multi-conv rl-based memory agent. Preprint, arXiv:2507.02259. — [arXiv:2507.02259](https://arxiv.org/abs/2507.02259) find
  45. Guibin Zhang, Hejia Geng, Xiaohang Yu, Zhenfei Yin, Zaibin Zhang, Zelin Tan, Heng Zhou, Zhongzhi Li, Xiangyuan Xue, Yijiang Li, Yifan Zhou, Yang Chen, Chen Zhang, Yutao Fan, Zihu Wang, Songtao Huang, Yue Liao, Hongru Wang, Mengyue Yang, Heng Ji, Michael Littman, Jun Wang, Shuicheng Yan, Philip Torr,… — [arXiv:2509.02547](https://arxiv.org/abs/2509.02547) find
  46. Zeyu Zhang, Quanyu Dai, Xiaohe Bo, Chen Ma, Rui Li, Xu Chen, Jieming Zhu, Zhenhua Dong, and Ji-Rong Wen. 2025b. A survey on the memory mechanism of large language model-based agents. ACM Trans. Inf. Syst., 43(6):155:1–155:47. find
  47. Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, Yifan Du, Chen Yang, Yushuo Chen, Zhipeng Chen, Jinhao Jiang, Ruiyang Ren, Yifan Li, Xinyu Tang, Zikang Liu, Peiyu Liu, Jian-Yun Nie, and Ji-Rong Wen. 2024. A survey of l…
  48. Wanjun Zhong, Lianghong Guo, Qiqi Gao, He Ye, and Yanlin Wang. 2024. Memorybank: Enhancing large language models with long-term memory. In Thirty-Eighth AAAI Conference on Artificial Intelligence, AAAI 2024, Thirty-Sixth Conference on Innovative Applications of Artificial Intelligence, IAAI 2024, Fo… find
  49. A gering memory management once the accumulated context reaches the budget, while baseline methods retain their original context limits (e.g., 128K for DeepResearch-30B-A3B). We report Pass@1 as the evaluation metric and use GPT-4o-2024-08-06 as the LLM-based judge. All experiments are conducted on… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/qian-2026-memobrain-executive-memory-as-agentic/