OpenAI’s AI Model Disproves Century-Old Geometry Conjecture, Marking a Milestone in AI Reasoning

OpenAI has announced that its latest reasoning model has produced an original mathematical proof disproving a long-standing conjecture in geometry first posed by the renowned mathematician Paul Erdős in 1946. This development marks a significant leap in AI’s ability to engage in complex, abstract reasoning and has sparked both excitement and scrutiny within the scientific community.

The Erdős Conjecture and Its Disproof

The conjecture in question, related to the maximum number of unit distances between points in a plane, had remained unsolved for nearly 80 years. For decades, mathematicians believed the optimal solution resembled a square grid layout. However, OpenAI’s model has uncovered a new family of constructions that outperform this traditional approach.

“For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids,” OpenAI stated in a post on X. “An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.”

Context: A History of AI and Mathematical Claims

This is not the first time OpenAI has made headlines for its AI’s mathematical capabilities. In 2025, former VP Kevin Weil claimed that GPT-5 had solved 10 previously unsolved Erdős problems. However, subsequent analysis revealed that the model had merely rediscovered existing solutions in the literature, leading to criticism from rivals like Yann LeCun and Google DeepMind CEO Demis Hassabis. Weil later removed his post.

OpenAI’s current claim, however, is supported by companion remarks from mathematicians Noga Alon, Melanie Wood, and Thomas Bloom, who maintains the Erdos Problems website. Bloom previously criticized Weil’s earlier claims as “a dramatic misrepresentation,” but he has since acknowledged the validity of OpenAI’s latest work.

Implications for AI and Mathematics

OpenAI emphasizes that this achievement is not the result of a specialized math-solving system but rather a general-purpose reasoning model. The company argues that this demonstrates AI’s growing capacity to handle long, intricate chains of reasoning and connect ideas across disciplines.

“AI is helping us to more fully explore the cathedral of mathematics we have built over the centuries,” Bloom said in a statement. “What other unseen wonders are waiting in the wings?”

The implications of this breakthrough extend beyond pure mathematics. OpenAI suggests that such advancements could revolutionize fields like biology, physics, engineering, and medicine by enabling AI to tackle complex, interdisciplinary problems.

Expert Reactions and Future Prospects

The mathematical community remains divided. While some experts praise the achievement as a testament to AI’s evolving capabilities, others caution against overestimating the model’s autonomy. “This is a remarkable step, but it’s important to remember that the model’s success is built on decades of human mathematical research,” said Dr. Sarah Friar, CFO of OpenAI.

As AI continues to push the boundaries of what is computationally possible, the collaboration between human mathematicians and machine intelligence may redefine the future of scientific discovery. OpenAI’s latest feat underscores the potential for AI not just to assist in research but to contribute original insights that challenge long-held assumptions.

For now, the disproof of Erdős’ conjecture stands as a landmark moment in the intersection of artificial intelligence and mathematics—a reminder of both the power and the responsibility that comes with developing ever-more capable systems.