AI Cracks 80-Year-Old Math Problem That Stumped the World’s Greatest Mathematicians

In a development that has sent shockwaves through the scientific community, OpenAI announced that one of its general-purpose reasoning models has autonomously solved the planar unit distance problem, an 80-year-old mathematics conjecture that had resisted every attempt by the world’s finest mathematicians to crack it.

What Is the Planar Unit Distance Problem?

First posed by the legendary Hungarian mathematician Paul Erdos in the 1940s, the problem seems deceptively simple: given a set of points on a flat plane, what is the maximum number of pairs of points that can be exactly one unit apart? Erdos believed the answer lay in grid-like arrangements, where the number of such connections would barely exceed the number of points themselves. For decades, the best improvements to his conjecture came more than 40 years ago, until now.

How Did the AI Solve It?

OpenAI’s model took an entirely unexpected approach. It borrowed a sophisticated technique from algebraic number theory to construct vast, multidimensional lattices in far higher dimensions than the two we are used to. Once built, the AI collapsed these complex shapes back down to two dimensions, producing a shadow of the higher-dimensional structures, and in doing so, proved that Erdos was significantly wrong. Points arranged in less symmetric patterns can yield far more unit-distance pairs than anyone thought possible.

Reaction from the Mathematical Community

Tim Gowers, a Fields Medalist at the University of Cambridge, wrote that the solution is a milestone in AI mathematics. In a blog post accompanying the work, he said: If a human had written the paper and submitted it to the Annals of Mathematics and I had been asked for a quick opinion, I would have recommended acceptance without any hesitation. No previous AI-generated proof has come close to that.

Princeton mathematician Will Sawin admitted his immediate reaction was disbelief. He told New Scientist: I thought the way that it was trying to solve it would not work, but then I looked at it more and I convinced myself that it does work. This is the most significant achievement by AI in mathematics so far.

What This Means for the Future

The implications are enormous. An AI capable of original mathematical discovery could eventually unlock breakthroughs across science, engineering, and medicine. Unlike human mathematicians who tend to specialize in narrow fields, this AI demonstrated the ability to draw from multiple advanced disciplines, including algebraic number theory, geometry, and combinatorics, to find a solution no single human expert could have reached alone.

Kevin Buzzard, a mathematician at Imperial College London, noted: Like many other AI breakthroughs, it did not take humans long at all to internalize, understand and generalize the arguments. One can contrast this with some human breakthroughs which have taken the community months or years to validate.

The Bigger Picture: AI Historic Week

This breakthrough comes during an extraordinary week for the AI industry. OpenAI and Anthropic are racing toward public offerings that could value them in the trillions. Nvidia posted .6 billion in quarterly revenue as demand for AI hardware goes parabolic, according to CEO Jensen Huang. And President Trump convened top tech CEOs in Washington to discuss AI security frameworks, a sign that the technology rapid advance is drawing attention at the highest levels of government.

For mathematicians, the message is clear: the era of AI-assisted discovery is no longer a distant promise. It is here, and it has already outperformed the best human minds on a problem that stood unsolved for eight decades. The question now is not whether AI can contribute to mathematics, but how quickly it will redefine the boundaries of what is possible.