EurekAgent: Agent Environment Engineering is All You Need For Autonomous Scientific Discovery
2026-06-11 • Artificial Intelligence
Artificial IntelligenceComputation and Language
AI summaryⓘ
The authors explain that while AI agents based on large language models can help automate scientific discovery, the main challenge now is creating the right environments for these agents to work well. They built a system called EurekAgent that provides tools and rules to guide the agents, making them explore better, manage work carefully, and work with less costly mistakes. Their approach improves performance on several math and machine learning problems and uses less money to do so. They suggest future research should focus on designing these supportive environments to make AI-driven science more reliable.
LLM-based agentsautonomous scientific discoveryenvironment engineeringpermissions engineeringartifact engineeringbudget engineeringhuman-in-the-loop26-circle packingAPI cost
Authors
Amy Xin, Jiening Siow, Junjie Wang, Zijun Yao, Fanjin Zhang, Jian Song, Lei Hou, Juanzi Li
Abstract
LLM-based agents have shown increasing potential in automating scientific discovery. Given an optimizable metric and an execution environment, they can propose, validate, and iterate scientific solutions, and have produced results that outperform human-designed approaches. As model capabilities continue to improve, we argue that the bottleneck for autonomous scientific discovery is shifting from prescribing agent workflows to designing agent environments: the resources, constraints, and interfaces that shape agent behavior. We frame this as environment engineering: building environments that amplify productive behaviors, such as open-ended exploration, systematic artifact management, and inter-agent collaboration, while suppressing harmful behaviors, such as reward hacking and high-friction human oversight. We present EurekAgent, an environment-engineered agent system for metric-driven autonomous scientific discovery. EurekAgent engineers the environment along four dimensions: permissions engineering for bounded agent execution and isolated evaluation; artifact engineering for filesystem and Git-based collaboration; budget engineering for budget-aware exploration; and human-in-the-loop engineering for easy human supervision and intervention. EurekAgent sets new state-of-the-art results on multiple mathematics, kernel engineering, and machine learning tasks, including new state-of-the-art 26-circle packing results discovered with less than $11 in total API cost. We open-source our code and results, and call for environment engineering as a core research direction for developing reliable autonomous research agents.