The paper introduces OpenConjecture, a dataset of mathematical conjectures from arXiv papers since February 2026, and evaluates frontier GPT models on 200 randomly sampled conjectures. The models generate candidate solutions for about 60% of the conjectures, with counterexamples occurring roughly 1.3 times more often than proofs. The study examines solution novelty, citation practices, and fidelity to the original statements, finding that some solutions deviate from authors’ intent and that many proofs closely follow existing results, though some exhibit moderate novelty.
By Helen Jenne, Davis Brown, Jesse He, Henry Kvinge
The article audits 18 chapter‑specific reviews of ten mathematical results announced by OpenAI, examining review standards, Lean formalizations, subsequent research, and references. No confirmed substantive errors remain in the principal results, though review depth varies and some dependencies are only partially checked. Chapter 8 raises the strongest reservation, requiring major revision of compressed analytic arguments, while Chapter 6’s polarity error was corrected after a PDF extraction issue.
By Miko{\l}aj Sienicki, Krzysztof Sienicki
The paper introduces GAP, a methodology that automatically generates mathematically equivalent variants of existing math problems using surface renames and kernel rewrites. Applying GAP to all 1,051 William Lowell Putnam Competition problems creates PutnamGAP, a 6,306-item dataset with 5,255 unseen variants. Evaluation of 18 commercial and open-source LLMs shows accuracy drops across all models, especially under kernel rewrites, indicating a persistent weakness in transferring proof plans to new mathematical settings.
By Yuren Hao, Xiang Wan, ChengXiang Zhai
arXiv:2606. 08728v1 Announce Type: new Abstract: Mathematical reasoning has long served as a stringent test of machine intelligence; over the past decade, it has moved from a niche problem within NLP to one of the most consequential AI frontiers.
By Syed Rifat Raiyan, Mohsinul Kabir, Hasan Mahmud, Md Kamrul Hasan
The paper announces a new lower bound of 0.8559 for the Steiner ratio, improving on the previous 0.824 bound for the Gilbert‑Pollak Conjecture. It introduces an AI system that uses large language models to generate rule‑constrained geometric lemmas, which are then turned into executable verification functions that certify the bound. The approach relies on only thousands of LLM calls, highlighting the feasibility of LLM‑based methods for advanced mathematical research.
By Yisi Ke, Tianyu Huang, Yankai Shu, Di He, Jingchu Gai, Liwei Wang
The paper introduces a method for evaluating the intrinsic interestingness of mathematical theorems by comparing the length of their proofs to the length of their statements. It trains a 27B language model to predict proof difficulty, enabling the generation and selection of more interesting theorems while significantly reducing overlap with existing Mathlib. The approach allows iterative expansion of a self‑building, machine‑verified mathematical library guided by quantifiable metrics.
By Niket Patel, Ahmad Rammal, Amaury Hayat, Remi Munos, Julia Kempe
Automatically constructing well-specified and valuable mathematical conjectures remains a central challenge in AI-assisted mathematical discovery. Many existing open problems and conjectures are often too broad, underspecified, or difficult to connect to plausible proof or refutation strategies.
The paper introduces HorizonMath, a benchmark of 113 largely unsolved mathematical problems across eight domains, paired with an open-source framework for automated verification. It focuses on the generator‑verifier gap, targeting problems that are hard to discover but easy to verify computationally, thereby avoiding costly formal proof verification or manual review. Using this framework, the authors found six novel solutions—three each from GPT‑5.4 Pro and GPT‑5.6 Sol—demonstrating that current models can contribute to mathematical research, while most state‑of‑the‑art models score below 10%.
By Erik Y. Wang, Sumeet R. Motwani, James V. Roggeveen, Eliot Hodges, Dulhan Jayalath, Charles London, Kalyan Ramakrishnan, Jakob Foerster, Cheng Zhang, Flaviu Cipcigan, Philip Torr, Alessandro Abate
The article reports an independent human audit of a proof from OpenAI’s "Ten Advances in Mathematics and Theoretical Computer Science." An initial review flagged a polarity error in a greedy conditioning lemma, but this was traced to a PDF text‑extraction issue that removed an overbar from a symbol. After correcting the rendering, no mathematical error was found, and the audit underscores that errors in AI‑generated mathematics can stem from document‑processing problems as well as reasoning mistakes.
By Miko{\l}aj Sienicki, Krzysztof Sienicki
arXiv:2608. 11941v1 Announce Type: new Abstract: We construct OEIS Open, a benchmark based on 492 open mathematical conjectures from the OEIS, formalized in Lean by Tsoukalas et al.
By Tom Adamczewski
The paper introduces a framework that uses large language models to represent and measure mathematical arguments, capturing statements, equations, assumptions, and constructions while distinguishing source occurrences from contextual states and inference records. It defines a directed acyclic graph representation where depth is the longest prerequisite path and bulk counts distinct ancestor edges, providing a recurrence that handles shared subarguments and preserves macro skeletons. The framework includes a standardized, versioned prompt protocol for annotation discovery, deterministic recomputation, and review assistance, and proposes cross‑paper links to study theorem co‑use and support probabilistic premise suggestions and review of omissions.
By Eunji Park
The paper introduces a new human‑AI collaboration paradigm for mathematical discovery, shifting from selecting individual problems to exploring broad research directions. It presents the Find, Attempt, and Recommend (FAR) pipeline, which automatically searches a literature corpus, filters candidate conjectures, and surfaces promising resolutions for expert review. In a combinatorics pilot, FAR processed over 5,000 papers, identified thousands of open conjectures, and ultimately highlighted 77 items that led to new discoveries.
By Zeyu Zheng, Shengtong Zhang, Jeremy Avigad, Prasad Tetali, Sean Welleck