Lune

ICML2026Top-tier venue

Numina-Lean-Agent: An Open and General Agentic Reasoning System for Formal Mathematics

Junqi Liu, Zihao Zhou, Zekai Zhu, Marco Dos Santos, Weikun He, jiawei liu, Yunzhou Xie, Junqiao Zhao, Qiufeng Wang, Lihong Zhi, Jia LI, Wenda Li

Abstract

Agentic systems have recently become the dominant paradigm for formal theorem proving, achieving strong performance by coordinating multiple models and tools. However, existing approaches often rely on task-specific pipelines and trained formal provers, limiting their flexibility and reproducibility. In this paper, we propose the paradigm that directly uses a general coding agent as a formal math reasoner. This paradigm is motivated by (1) A general coding agent provides a natural interface for diverse reasoning tasks beyond proving, (2) Performance can be improved by simply replacing the underlying base model, without training., and (3) MCP enables flexible extension and autonomous calling of specialized tools, avoiding complex design. Based on this paradigm, we introduce Numina-Lean-Agent, which combines Claude Code with Numina-Lean-MCP to enable autonomous interaction with Lean, retrieval of relevant theorems, informal proving and auxiliary reasoning tools. Using Claude Opus 4.5 as the base model, Numina-Lean-Agent solves all problems in Putnam 2025 (12/12), matching the best closed-source system. Beyond benchmark evaluation, we further demonstrate its generality by interacting with mathematicians to successfully formalize the Brascamp-Lieb theorem. We release Numina-Lean-Agent and all solutions at https://github.com/project-numina/numina-lean-agent .

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

1 top-tier paper has cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • miniF2F-Dafny: LLM-Guided Mathematical Theorem Proving via Auto-Active Verification cites HyperTree Proof Search for Neural Theorem Proving
  • DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search cites HyperTree Proof Search for Neural Theorem Proving
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers1

  1. miniF2F-Dafny: LLM-Guided Mathematical Theorem Proving via Auto-Active VerificationMantas Baksys, Stefan Zetzsche, Olivier Bouissou, Sean B HoldenICML 2026 · 2 citations

Builds on2

  1. HyperTree Proof Search for Neural Theorem ProvingGuillaume Lample, Timothée Lacroix, Marie-Anne Lachaux, Aurélien Rodriguez et al.NeurIPS 2022 · 271 citations
  2. DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree SearchHuajian Xin, Z. Z. Ren, Junxiao Song, Zhihong Shao et al.ICLR 2025

Related papers

  1. LeanAgent: Lifelong Learning for Formal Theorem ProvingAdarsh Kumarappan, Mo Tiwari, Peiyang Song, Robert Joseph George et al.ICLR 2025
  2. HERMES: Towards Efficient and Verifiable Mathematical Reasoning in LLMsAzim Ospanov, Zijin Feng, Jiacheng Sun, Haoli Bai et al.ICML 2026 · 5 citations
  3. APOLLO: Automated LLM and Lean Collaboration for Advanced Formal ReasoningAzim Ospanov, Farzan Farnia, Roozbeh MohitNeurIPS 2025 · 49 citations
  4. Hilbert: Recursively Building Formal Proofs with Informal ReasoningSumanth Varambally, Thomas Voice, Yanchao Sun, Zhifeng Chen et al.ICLR 2026 · 62 citations
  5. ProofOptimizer: Training Language Models to Simplify Proofs without Human DemonstrationsAlex Gu, Bartosz Piotrowski, Fabian Gloeckle, Kaiyu Yang et al.ICLR 2026 · 11 citations
Dusk over the sea between two cliffs drawn in fine vertical lines