An OpenAI model has disproved a central conjecture in discrete geometry

2026-05-12 · Source: OpenAI News · Field: Science & Research — Mathematics & Computational Sciences, Artificial Intelligence & Machine Learning, Research Methodology & Innovation · Depth: Expert, medium

Summary

An internal OpenAI model has disproved a central conjecture in discrete geometry, specifically the planar unit distance problem posed by Paul Erdős in 1946. For nearly 80 years, mathematicians believed that square grid constructions were optimal for maximizing unit-distance pairs among $n$ points in a plane, yielding $n^{1 + C / \log \log(n)}$ pairs. The OpenAI model, a new general-purpose reasoning model, provided an infinite family of examples demonstrating at least $n^{1+\delta}$ unit-distance pairs for a fixed exponent $\delta > 0$, with a forthcoming refinement suggesting $\delta=0.014$. This marks the first time an AI has autonomously solved a prominent open problem in mathematics, utilizing unexpected sophisticated ideas from algebraic number theory, such as infinite class field towers and Golod–Shafarevich theory, to construct the counterexample.

Key takeaway

For AI Scientists and Research Scientists evaluating the capabilities of advanced reasoning models, this breakthrough demonstrates that AI can autonomously solve prominent, long-standing mathematical problems. Your teams should consider integrating general-purpose AI models into frontier research efforts, especially for problems requiring deep reasoning and the synthesis of ideas from distant knowledge domains, to potentially uncover novel solutions and accelerate discovery.

Key insights

An OpenAI model disproved a long-standing mathematical conjecture using algebraic number theory, marking a significant AI reasoning milestone.

Principles

• AI can autonomously resolve open mathematical problems.
• Sophisticated reasoning models can connect disparate mathematical fields.

Method

The AI model generated a proof by replacing Gaussian integers with more complex generalizations from algebraic number theory, leveraging concepts like infinite class field towers and Golod–Shafarevich theory to construct configurations with more unit-distance pairs.

In practice

• Explore AI for complex problem-solving in STEM.
• Investigate cross-disciplinary connections revealed by AI.

Best for: AI Scientist, Research Scientist, Director of AI/ML

Related on AIssential

• OpenAI’s math breakthrough played to AI’s strengths — AI disproved an 80-year-old math conjecture by applying broad knowledge and persistent strategy exploration, marking a significant autonomous contribution.
• OpenAI claims it solved an 80-year-old math problem — for real this time — OpenAI's new model disproved an 80-year-old math conjecture, demonstrating advanced autonomous reasoning in AI.
• OpenAI shifts the boundary of automated reasoning with a "milestone in AI mathematics" that experts are now unpacking — OpenAI's AI disproved a long-standing mathematical conjecture by applying unexpected tools from algebraic number theory.
• 😸 OpenAI solved an 80-year math problem by... disproving it — An OpenAI model disproved an 80-year math conjecture, showcasing AI's advanced reasoning and verifiable problem-solving capabilities.
• AI cracks 80-year-old mathematics challenge — researchers are astonished — AI can autonomously solve complex, long-standing mathematical conjectures.
• An OpenAI model has disproved a central conjecture in discrete geometry — An OpenAI model resolved Erdos Problem 90, demonstrating AI's capacity to solve complex, long-unsolved mathematical conjectures and sparking debate on AI novelty.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by OpenAI News.