Large language models (LLMs) are increasingly used in workflows for generating formal proofs in Lean. These workflows often decompose problems into smaller lemmas, sample many proof attempts, and use compiler feedback to guide search. However, they can be prohibitively expensive, often spending substantial compute on attempts that ultimately fail. In this work, we address this problem with an action routing agent that consists of a data plane and a control plane. The data plane generates natural-language lemma decompositions, formalizes them in Lean, and samples proof attempts for the resulting theorem and lemma targets. The control plane observes previous failed Lean attempts, estimates both the likelihood of success and the cost of another attempt, and decides whether to continue proving the current target or restart from a new breakdown. On a subset of PutnamBench, our agent decreases the cost by 28.9% over a fixed-step baseline on average, preserving performance while using substantially less compute. These results suggest that failed Lean trajectories provide actionable signals for cost-aware resource allocation in agentic theorem proving.
Discover and Prove: An Open-source Agentic Framework for Hard Mode Automated Theorem Proving in Lean 4
State Key Laboratory for Multimedia Information Processing, School of Computer Science, PKU-Anker LLM Lab, Peking University(信息处理国家重点实验室,计算机科学学院,PKU-Anker LLM实验室,北京大学);Huawei Technologies Co., Ltd.(华为技术有限公司);School of Software & Microelectronics, Peking University(软件与微电子学院,北京大学);School of Electronics Engineering and Computer Science, Peking University(电子工程与计算机科学学院,北京大学)
FMC: Formalization of Natural Language Mathematical Competition Problems
State Key Laboratory for Multimedia Information Processing, School of Computer Science, PKU-Anker LLM Lab, Peking University, Beijing, China(多媒体信息处理国家重点实验室,计算机科学学院,PKU-Anker LLM实验室,北京大学,北京,中国);Paul G. Allen School of Computer Science(保罗·G·阿伦计算机科学学校);Engineering, University of Washington, U.S.A.(工程,华盛顿大学,美国)
RePro: Proof-Verified Benchmark Rewriting for Reliable Evaluation of LLM Mathematical Problem Solving
Nanyang Technological University(南洋理工大学);The Chinese University of Hong Kong, Shenzhen(香港中文大学(深圳));INSAIT;Sofia University “St. Kliment Ohridski”(索非亚大学“圣·克利门特·奥赫里德斯基”);Shenzhen Loop Area Institute(深圳河套学院);AIRS
LeanMarathon: Toward Reliable AI Co-Mathematicians through Long-Horizon Lean Autoformalization
Department of Statistics, University of Warwick, UK(英国沃里克大学统计系);Center for Advanced Intelligence Project, RIKEN, Japan(日本理化学研究所高级智能项目);Department of Statistics, University of Michigan, USA(美国密歇根大学统计系);Department of Mathematical Informatics, The University of Tokyo(东京大学数学信息学系;日本理化学研究所高级智能项目);also Center for Advanced Intelligence Project, RIKEN, Japan(加州大学伯克利分校电气工程与计算机科学系;统计系);Department of Electrical Engineering and Computer Sciences, also Department of Statistics, University of California, Berkeley, USA(上海交通大学数学科学学院,自然科学院和MOE-LSC);School of Mathematical Sciences, Institute of Natural Sciences and MOE-LSC, Shanghai Jiao Tong University, China