OpenAI revealed Monday that its general-purpose reasoning model has solved the planar unit distance problem, an open question in discrete geometry first posed by mathematician Paul Erdos in 1946. The achievement marks the first time an AI system has autonomously cracked a prominent open problem central to a field of mathematics.
What the AI discovered
For nearly 80 years, mathematicians believed optimal solutions to the Erdos problem roughly resembled square grids. The OpenAI model disproved that assumption, generating an entirely new family of constructions that delivers better performance. The proof came from a general-purpose reasoning model, not a system built specifically for mathematics, and was formalized in the Lean verification language.
Expert verification
Fields Medalist Terence Tao reviewed and accepted the proof, lending it credibility within the mathematical community. The result has been described as the beginning of a new era where AI systems can hold together long chains of reasoning and connect ideas across remote fields.
Implications beyond math
OpenAI said the same capabilities that enabled this mathematical breakthrough will soon accelerate work in biology, physics, engineering, and medicine. The company emphasized that human judgment remains essential — AI can search, suggest, and verify, but people choose the problems that matter and interpret results.