
Share
A new AI model from OpenAI has cracked a nearly 80-year-old math problem, challenging the assumptions about the limits of machine reasoning and opening new avenues for algorithmic breakthroughs.
For nearly eight decades, mathematicians have grappled with a deceptively simple question in combinatorial geometry: if you place ( n ) points in the plane, how many pairs of points can be exactly one unit apart? This is known as the planar unit distance problem, first posed by Paul Erdős in 1946. Despite its simplicity, it has remained a significant challenge, with even Erdős offering a monetary prize for its resolution.
Today, OpenAI has announced a breakthrough that disproves a longstanding conjecture about this problem. The prevailing belief since Erdős’s original work was that the "square grid" constructions were essentially optimal for maximizing the number of unit-distance pairs. However, an internal OpenAI model has now provided an infinite family of examples that yield a polynomial improvement over these grids.
The proof is notable not only for its mathematical content but also for how it was discovered. Unlike previous AI systems that were trained specifically for mathematics or scaffolded to search through proof strategies, this breakthrough came from a new general-purpose reasoning model. This model, part of OpenAI’s broader effort to test the capabilities of advanced models in frontier research, was evaluated on a collection of Erdős problems and managed to produce a proof resolving this open problem.

Mathematical Impact: The result is significant because it marks the first time an AI has autonomously solved a prominent open problem central to a subfield of mathematics. It demonstrates the depth of reasoning these systems can now support, where each step in a long argument must hold together from beginning to end.
AI Research Milestone: Fields medalist Tim Gowers, writing in the companion paper, calls the result "a milestone in AI." The proof brings sophisticated ideas from algebraic number theory to bear on an elementary geometric question, showcasing the model’s ability to integrate advanced mathematical concepts.
Disproving Conjectures: OpenAI's model has disproved a long-standing conjecture about the planar unit distance problem, providing new constructions that yield a polynomial improvement over previous best-known results.
General-Purpose Reasoning: The breakthrough comes from a general-purpose reasoning model, not one specifically trained for mathematics. This suggests that advanced AI models can contribute to cutting-edge research in various fields.
Mathematical and AI Collaboration: The proof has been verified by external mathematicians, highlighting the potential for collaboration between human experts and AI systems in solving complex problems.
This breakthrough is a testament to the growing capabilities of AI in reasoning and problem-solving. As companies like Balyasny Asset Management leverage advanced models for investment research and tech giants like Microsoft and Google compete in developing AI coding tools, the implications of this proof extend beyond mathematics. It signals a new era where AI can autonomously contribute to solving some of the most challenging problems across various domains.
Tags
Original Sources
An OpenAI model has disproved a central conjecture in discrete geometry
↗ https://links.tldrnewsletter.com/Ki0Bx1
About the author
Kai built ML infrastructure at a Bay Area startup before developing an obsession with transformer architectures and inference optimisation that eventually pulled him out of product work entirely. A stint at a compute research lab sharpened his instinct for what actually matters in a model release versus what is marketing. He writes from the inside — from the perspective of someone who has debugged the systems he is describing at three in the morning. He is allergic to hype and instinctively drawn to the unglamorous plumbing questions that everyone else skips over.
More from The Engineer →This Week's Edition
8 June 2026
67 articles
Related Articles

Global Researchers Compete to Shape AI's Future in Organizations
Models & Research · 3 min

MiniMax M3 Challenges GPT-5.5 and Gemini 3.1 Pro with Superior Performance at a Fraction of the Cost
Models & Research · 4 min

OpenAI’s AI Solves Complex Math Problems by Leveraging Pattern Recognition
Models & Research · 4 min
Related Articles

Global Researchers Compete to Shape AI's Future in Organizations
Models & Research · 3 min

MiniMax M3 Challenges GPT-5.5 and Gemini 3.1 Pro with Superior Performance at a Fraction of the Cost
Models & Research · 4 min

OpenAI’s AI Solves Complex Math Problems by Leveraging Pattern Recognition
Models & Research · 4 min
More Stories
© 2026 Cedar & Bloom. All rights reserved.