导读
英文标题：FrontierPhysics: Evaluating Agents for End-to-End Physics Research arXiv编号：2611.06910 作者：Bingran You, Wenjun Ke, Ria Rosenauer 来源：arXiv 原文链接：http…
【编辑部】这对你意味着什么
这篇论文介绍了一个名为 FrontierPhysics 的新基准，用于评估 AI agent 完成端到端物理研究的能力。基准包含 56 个任务，由 42 位物理学家基于自身研究创建，涵盖理论、计算和实验物理九大领域。测试了 22 个模型的 26 种配置，没有任何配置能通过超过 46 次（共 168 次）测试。专家评审指出，当前前沿 agent 的主要短板在于推理不够严谨和缺乏专家直觉，而非物理知识或执行错误。对于从事物理或相关 AI 研究的华人学者和学生，这提供了一个更贴近真实科研流程的评测标准。
英文标题：FrontierPhysics: Evaluating Agents for End-to-End Physics Research
arXiv编号：2611.06910
作者：Bingran You, Wenjun Ke, Ria Rosenauer
来源：arXiv
原文链接：https://arxiv.org/abs/2611.06910
中文摘要：
AI agents are playing an increasingly important role in physics research. However, existing physics benchmarks only evaluate fixed questions, isolated sub-tasks, or paper reproduction with given solutions, and rarely check research quality alongside correctness. We introduce FrontierPhysics, an end-to-end benchmark with 56 tasks made by 42 physicists based on their own research, covering nine domains across theoretical, computational, and experimental physics. Each task requires agents to review the literature, develop a research plan, conduct research and experiments, and produce structured outputs and scientific manuscripts. Combining deterministic verifiers with expert-authored rubrics, FrontierPhysics evaluates whether agents' research is correct and publishable. Across 22 models in 26 configurations, no configuration passes more than 46 of its 168 rollouts. Manual review of agent trajectories shows that frontier agents' main weaknesses are insufficiently rigorous reasoning and limited expert intuition, rather than gaps in physics knowledge or execution errors.
本文由华人空间 AI 论文助手自动抓取翻译，仅供学术交流。
【编辑部】今日讨论
你在科研中使用 AI agent 时，更常遇到推理不严谨的问题，还是缺乏领域直觉的问题？
