Elamur Research⌕
☰
home / papers / 2026 / CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery

CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery

Paper · 2026 · Ao Qu, Han Zheng, Zijian Zhou et al.

Summary

LLM-driven evolutionary systems such as FunSearch and AlphaEvolve search for better programs on problems where the optimum is unknown. However, a hard-coded algorithm, not the model, decides which earlier solutions to build on, when to test, and what to remember. CORAL hands these decisions to long-running coding agents. The agents work asynchronously in isolated workspaces and share a persistent memory of scored attempts, notes, and reusable skills. Periodic "heartbeat" prompts push them to reflect or change direction. Evaluators run separately so agents cannot game the scoring.

The authors tested CORAL on mathematical, algorithmic, and systems-optimization tasks, using Claude Code with Opus 4.6. It beat fixed evolutionary baselines such as OpenEvolve, reaching new state-of-the-art results while needing far fewer evaluations. On Anthropic's GPU kernel benchmark, four cooperating agents lowered the best known result from 1363 to 1103 cycles, roughly a 20% gain.

Ablations show that both accumulated knowledge and genuine agent-to-agent co-evolution drive the gains, not just extra compute. This suggests autonomous multi-agent search is a strong direction for open-ended discovery.

Key points

  • Headline results: The abstract reports new SOTA on 10 tasks, with 3–10× higher improvement rates than fixed evolutionary search. The introduction, apparently from an earlier version, says 8 of 11 tasks, a 2.5× improvement rate, and 10× fewer evaluations. Code is at github.com/Human-Agent-Society/CORAL.
  • Kernel Engineering task (baseline 147,734 cycles):
    • OpenEvolve: 2740 cycles, 3% of attempts improved.
    • CORAL with 1 agent: 1350 cycles, 43% improved.
    • CORAL with 4 agents: 1103 cycles, a 133.9× speedup.
  • Knowledge ablation: Turning off note and skill creation raised the single-agent kernel result from 1350 to 1601 cycles, an 18.6% regression.
  • Co-evolution vs. independent runs: Four co-evolving agents beat the best of four independent runs on all three stress tasks:
    • Kernel Engineering: 1103 vs. 1180 cycles.
    • Polyominoes: 84.2 vs. 80.8.
    • Transaction Scheduling: 4694 vs. 4629.
  • Cross-agent reuse on Kernel Engineering: 36% of attempts started from another agent's commit. These improved at 17%, versus 9% across all attempts, and produced 66% of new records.

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

🎓Stanford UniversityStanford · United States567 works · 335,190 citations🎓National University of SingaporeSingapore · Singapore221 works · 39,953 citations🏢AmazonSeattle · United States131 works · 41,784 citations🎓McGill UniversityMontreal · Canada97 works · 11,755 citations

Also on the byline: Singapore-MIT Alliance for Research and Technology · MiniMax (People's Republic of China).

 Fields, topics and concepts

📚 Shelves 2The field of science this work belongs to (OpenAlex fields).Computer Science · 19,503 worksArtificial Intelligence · 18,805 works

Cite

Cite as Ao Qu, Han Zheng, Zijian Zhou, Yihao Yan, Yihong Tang, Shao Yong Ong, Fenglu Hong, Kaichen Zhou, Chonghe Jiang, Minwei Kong, Jiacheng Zhu, Xuan Jiang, Sirui Li, Cathy Wu, Bryan Kian Hsiang Low, Jinhua Zhao and Paul Pu Liang. 2026. CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery. arXiv:2604.01658.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.2604.01658 · arXiv 2604.01658

@misc{qu2026coral,
  title = {{CORAL: Towards Autonomous Multi-Agent Evolution for Open-Ended Discovery}},
  author = {Ao Qu, Han Zheng, Zijian Zhou, Yihao Yan, Yihong Tang, Shao Yong Ong, Fenglu Hong, Kaichen Zhou, Chonghe Jiang, Minwei Kong, Jiacheng Zhu, Xuan Jiang, Sirui Li, Cathy Wu, Bryan Kian Hsiang Low, Jinhua Zhao and Paul Pu Liang},
  year = {2026},
  doi = {10.48550/arxiv.2604.01658},
  eprint = {2604.01658},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2604.01658}
}

References

44 references, 16 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Lakshya A Agrawal, Shangyin Tan, Dilara Soylu, Noah Ziems, Rishi Khare, Krista Opsahl-Ong, Arnav Singhvi, Herumb Shandilya, Michael J Ryan, Meng Jiang, Christopher Potts, Koushik Sen, Alexandros G. Dimakis, Ion Stoica, Dan Klein, Matei Zaharia, and Omar Khattab. GEPA: Reflective prompt evolution can… find
  2. Anthropic. Performance take-home assignment, 2025. https://github.com/anthropics/ original performance takehome. ↗
  3. Gustavo Assumpção et al. CodeEvolve: An open source evolutionary coding agent for algorithm discovery and optimization. arXiv preprint, 2026. find
  4. Mert Cemri, Shu Liu, Shubham Agarwal, Monishwaran Maheswaran, Zhifei Li, Qiuyang Mang, Ashwin Naren, Kurt Keutzer, Alexandros G. Dimakis, Koushik Sen, Matei Zaharia, and Ion Stoica. AdaEvolve: Adaptive LLM driven zeroth-order optimization. arXiv preprint arXiv:2602.20133, 2026. find
  5. Hongzheng Chen, Yingheng Wang, Yaohui Cai, Hins Hu, Jiajie Li, Shirley Huang, Chenhui Deng, Rongjian Liang, Shufeng Kong, Haoxing Ren, et al. Heurigym: An agentic benchmark for llm-crafted heuristics in combinatorial optimization. arXiv preprint arXiv:2506.07972, 2025. — [arXiv:2506.07972](https://arxiv.org/abs/2506.07972) find
  6. Terry Chen, Zhifan Ye, Bing Xu, Zihao Ye, Timmy Liu, Ali Hassani, Tianqi Chen, Andrew Kerr, Haicheng Wu, Yang Xu, Yu-Jung Chen, Hanfeng Chen, Aditya Kane, Ronny Krashinsky, Ming-Yu Liu, Vinod Grover, Luis Ceze, Roger Bringmann, John Tran, Wei Liu, Fung Xie, Michael Lightstone, and Humphrey Shi. Avo:… find
  7. Weize Chen, Yusheng Su, Jingwei Zuo, Cheng Yang, Chenfei Yuan, Chi-Min Chan, Heyang Yu, Yaxi Lu, Yi-Hsin Hung, Chen Qian, Yujia Qin, Xin Cong, Ruobing Xie, Zhiyuan Liu, Maosong Sun, and Jie Zhou. AgentVerse: Facilitating multi-agent collaboration and exploring emergent behaviors. arXiv preprint arXi…
  8. Shuangrui Ding, Xuanlang Dai, Long Xing, Shengyuan Ding, Ziyu Liu, Yang JingYi, Penghui Yang, Zhixiong Zhang, Xilin Wei, Xinyu Fang, Yubo Ma, Haodong Duan, Jing Shao, Jiaqi Wang, Dahua Lin, Kai Chen, and Yuhang Zang. Wildclawbench: A benchmark for real-world, long-horizon agent evaluation, 2026. URL… — [arXiv:2605.10912](https://arxiv.org/abs/2605.10912) find
  9. Haoyuan Gao et al. A survey of self-evolving agents: On path to artificial super intelligence. arXiv preprint arXiv:2507.21046, 2025.
  10. Juraj Gottweis, Wei-Hung Weng, Alexander Daryin, Tao Tu, Anil Palepu, Petar Sirkovic, Artiom Myaskovsky, Felix Weissenberger, Keran Rong, Ryutaro Tanno, et al. Towards an ai co-scientist. arXiv preprint arXiv:2502.18864, 2025.
  11. Sirui Hong, Mingchen Zhuge, Jonathan Chen, Xiawu Zheng, Yuheng Cheng, Ceyao Zhang, Jinlin Wang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, Liyang Zhou, Chenyu Ran, Lingfeng Xiao, Chenglin Wu, and Jürgen Schmidhuber. MetaGPT: Meta programming for a multi-agent collaborative framework. arXiv preprin…
  12. Andrej Karpathy. autoresearch: Ai agents running research on single-gpu nanochat training automatically. https://github.com/karpathy/autoresearch, March 2026. GitHub repository. Accessed: 2026-03-31. ↗
  13. Valentin Khrulkov, Artem Galichin, Dmitry Bashkirov, Dmitry Vinichenko, Oleg Travkin, Roman Alferov, Artem Kuznetsov, and Ivan Oseledets. GigaEvo: An open source optimization framework powered by LLMs and evolution algorithms. arXiv preprint arXiv:2511.17592, 2025. — [arXiv:2511.17592](https://arxiv.org/abs/2511.17592) find
  14. Minwei Kong, Ao Qu, Xiaotong Guo, Wenbin Ouyang, Chonghe Jiang, Han Zheng, Yining Ma, Dingyi Zhuang, Yuhan Tang, Junyi Li, Shenhao Wang, Haris Koutsopoulos, Hai Wang, Cathy Wu, and Jinhua Zhao. AlphaOPT: Formulating optimization programs with self-improving LLM experience library. arXiv preprint arX… — [arXiv:2510.18428](https://arxiv.org/abs/2510.18428) find
  15. LangChain. LangGraph: Build resilient language agents as graphs, 2024. https://github.com/langchain-ai/langgraph. ↗
  16. Robert Tjarko Lange, Yuki Imajuku, and Edoardo Cetin. ShinkaEvolve: Towards open-ended and sample-efficient program evolution. arXiv preprint arXiv:2509.19349, 2025.
  17. Guohao Li, Hasan Hammoud, Hani Itani, Dmitrii Khizbullin, and Bernard Ghanem. Camel: Communicative agents for” mind” exploration of large language model society. Advances in neural information processing systems, 36:51991–52008, 2023. — CAMEL: Communicative Agents for "Mind" Exploration of Large Language Model Society
  18. Shu Liu, Shubham Agarwal, Monishwaran Maheswaran, Mert Cemri, Zhifei Li, Qiuyang Mang, Ashwin Naren, Ethan Boneh, Audrey Cheng, Melissa Z. Pan, Alexander Du, Kurt Keutzer, Alvin Cheung, Alexandros G. Dimakis, Koushik Sen, Matei Zaharia, and Ion Stoica. EvoX: Meta-evolution for automated discovery. a… — [arXiv:2602.23413](https://arxiv.org/abs/2602.23413) find
  19. Shu Liu, Mert Cemri, Shubham Agarwal, Alexander Krentsel, Ashwin Naren, Qiuyang Mang, Zhifei Li, Akshat Gupta, Monishwaran Maheswaran, Audrey Cheng, Melissa Pan, Ethan Boneh, Kannan Ramchandran, Koushik Sen, Alexandros G. Dimakis, Matei Zaharia, and Ion Stoica. Skydiscover: A flexible framework for… find
  20. Tengxiao Liu, Yuqing Yang, Xi Ye, and Danqi Chen. Can coding agents optimize algorithms autonomously? https://tengxiaoliu.github.io/autoevolver/, March 2026c. Project page. Accessed: 2026-04-01. ↗
  21. Chris Lu, Cong Lu, Robert Tjarko Lange, Jakob Foerster, Jeff Clune, and David Ha. The AI scientist: Towards fully automated open-ended scientific discovery. arXiv preprint arXiv:2408.06292, 2024.
  22. Aman Madaan, Niket Tandon, Prakhar Gupta, Skyler Hallinan, Luyu Gao, Sarah Wiegreffe, Uri Alon, Nouha Dziri, Shrimai Prabhumoye, Yiming Yang, Shashank Gupta, Bodhisattwa Prasad Majumder, Katherine Hermann, Sean Welleck, Amir Yazdanbakhsh, and Peter Clark. Self-refine: Iterative refinement with self-… find
  23. Qiuyang Mang, Wenhao Chai, Zhifei Li, Huanzhi Mao, Shang Zhou, Alexander Du, Hanchen Li, Shu Liu, Edwin Chen, Yichuan Wang, et al. Frontiercs: Evolving challenges for evolving intelligence. arXiv preprint arXiv:2512.15699, 2025. — [arXiv:2512.15699](https://arxiv.org/abs/2512.15699) find
  24. MiniMax. MiniMax M2.5: Built for real-world productivity, 2026. https://www.minimax.io/news/minimax-m25. ↗
  25. Alexander Novikov, Ngân Vũ, Marvin Eisenberger, Emilien Dupont, Po-Sen Huang, Adam Zsolt Wagner, Sergey Shirobokov, Borislav Kozlovskii, Francisco JR Ruiz, Abbas Mehrabian, et al. Alphaevolve: A coding agent for scientific and algorithmic discovery. arXiv preprint arXiv:2506.13131, 2025.
  26. OpenClaw Foundation. Openclaw: Your personal ai assistant, 2026. URL https://github.com/openclaw/openclaw. Open-source personal AI assistant. ↗
  27. OpenCode. OpenCode: The open source AI coding agent, 2025. https://opencode.ai. ↗
  28. Anne Ouyang, Simon Guo, Simran Arora, Alex L Zhang, William Hu, Christopher Re, and Azalia Mirhoseini. Kernelbench: Can llms write efficient gpu kernels? In Forty-second International Conference on Machine Learning, 2025a.
  29. Siru Ouyang, Jun Yan, I Hsu, Yanfei Chen, Ke Jiang, Zifeng Wang, Rujun Han, Long T Le, Samira Daruki, Xiangru Tang, et al. Reasoningbank: Scaling agent self-evolving with reasoning memory. arXiv preprint arXiv:2509.25140, 2025b. — [arXiv:2509.25140](https://arxiv.org/abs/2509.25140) find
  30. Chen Qian, Wei Liu, Hongzhang Liu, Nuo Chen, Yufan Dang, Jiahao Li, Cheng Yang, Weize Chen, Yusheng Su, Xin Cong, Juyuan Xu, Dahai Li, Zhiyuan Liu, and Maosong Sun. Chat-Dev: Communicative agents for software development. arXiv preprint arXiv:2307.07924, 2024. rllm-org. Hive. https://github.com/rllm… ↗
  31. Bernardino Romera-Paredes, Mohammadamin Barekatain, Alexander Novikov, Matej Balog, M. Pawan Kumar, Emilien Dupont, Francisco J. R. Ruiz, Jordan S. Ellenberg, Pengming Wang, Omar Fawzi, Pushmeet Kohli, and Alhussein Fawzi. Mathematical discoveries from program search with large language models. Natu…
  32. Aaditya Sharma. OpenEvolve: An open-source evolutionary coding agent. arXiv preprint, 2025. https://github.com/codelion/openevolve. ↗
  33. Noah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan, and Shunyu Yao. Reflexion: Language agents with verbal reinforcement learning. In Advances in Neural Information Processing Systems, volume 36, 2023.
  34. Xingyao Wang et al. OpenHands: An open platform for AI software developers as generalist agents. arXiv preprint arXiv:2407.16741, 2024.
  35. Yuxiang Wang, Shan Su, Zhiwei Zeng, Enhong Xu, Lei Ren, Xiao Yang, Zhi Huang, Xiangnan He, Liang Ma, Baolin Peng, Hao Cheng, Pengcheng He, Weizhu Chen, Shuming Wang, Simon S. Du, and Yelong Shen. ThetaEvolve: Test-time learning on open problems. arXiv preprint arXiv:2511.23473, 2025. — [arXiv:2511.23473](https://arxiv.org/abs/2511.23473) find
  36. Qingyun Wu, Gagan Bansal, Jieyu Zhang, Yiran Wu, Beibin Li, Erkang Zhu, Li Jiang, Xiaoyun Zhang, Shaokun Zhang, Jiale Liu, Ahmed Hassan Awadallah, Ryen W. White, Doug Burger, and Chi Wang. AutoGen: Enabling next-gen LLM applications via multi-agent conversation. arXiv preprint arXiv:2308.08155, 2023…
  37. Qingyun Wu, Gagan Bansal, Jieyu Zhang, Yiran Wu, Beibin Li, Erkang Zhu, Li Jiang, Xiaoyun Zhang, Shaokun Zhang, Jiale Liu, et al. Autogen: Enabling next-gen llm applications via multi-agent conversations. In First conference on language modeling, 2024. find
  38. Minhao Yan, Baolin Peng, Benjamin Coleman, Zihao Chen, Zhiwei Xie, Zhicheng He, Noveen Sachdeva, Ian Ye, Wei Wang, Chi Wang, Ed H. Chi, Wang Kang, Deborah Z. Cheng, and Bolin Wang. PACEvolve: Enabling long-horizon progress-aware consistent evolution. arXiv preprint arXiv:2601.10657, 2026. — [arXiv:2601.10657](https://arxiv.org/abs/2601.10657) find
  39. John Yang, Carlos E. Jimenez, Alexander Wettig, Kilian Liber, Shunyu Yao, Karthik Narasimhan, and Ofir Press. SWE-agent: Agent-computer interfaces enable automated software engineering. arXiv preprint arXiv:2405.15793, 2024.
  40. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. ReAct: Synergizing reasoning and acting in language models. In International Conference on Learning Representations, 2023.
  41. 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. Memagent: Reshaping long-context LLM with multi-conv RL-based memory agent. In The Fourteenth International Conference on Learning Representations, 202… find
  42. Mert Yuksekgonul, Daniel Koceja, Xiang Li, Federico Bianchi, Jack McCaleb, Xinyun Wang, Jan Kautz, Yejin Choi, James Zou, Carlos Guestrin, et al. Learning to discover at test time. arXiv preprint arXiv:2601.16175, 2026. find
  43. Han Zheng, Yining Ma, Brandon Araki, and Cathy Wu. Learning-guided prioritized planning for lifelong multi-agent path finding in warehouse automation. Journal of Artificial Intelligence Research, 85:1–30, 2026. doi: 10.1613/jair.1.20611. — [doi:10.1613/jair.1.20611](https://doi.org/10.1613/jair.1.20611) find
  44. Zijian Zhou, Ao Qu, Zhaoxuan Wu, Sunghwan Kim, Alok Prakash, Daniela Rus, Bryan Kian Hsiang Low, and Paul Pu Liang. MEM1: Learning to synergize memory and reasoning for efficient long-horizon agents. In The Fourteenth International Conference on Learning Representations, 2026. URL https://openreview… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/qu-2026-coral-towards-autonomous-multi-agent/