An OpenAI model solved a famous math problem that stumped humans for 80 years

2026-06-01 · Source: AI - Ars Technica · Field: Science & Research — Mathematics & Computational Sciences, Artificial Intelligence & Machine Learning · Depth: Intermediate, long

Summary

In mid-May, OpenAI announced that an internal AI model disproved the Erdős unit distance conjecture, a famous discrete geometry problem that had stumped mathematicians for 80 years. This marks a significant milestone, with Fields Medalist Tim Gowers calling it "a milestone in AI mathematics." The AI model achieved this by cleverly applying existing ideas, constructing a grid in high-dimensional space, and projecting it into two dimensions using algebraic integers, rather than pioneering new techniques. This approach demonstrated a more complex point arrangement that yields more unit-distance pairs than Erdős's square grid assumption, with human mathematician Will Sawin showing it grows at least at n^1.014. This achievement fits within a progression of AI in mathematics, from struggling with arithmetic to acing high school competitions and contributing to research in constrained settings.

Key takeaway

For research scientists evaluating AI's potential in mathematics, this development signals a shift towards AI models complementing human expertise. You should consider integrating AI systems to explore vast mathematical literature or to grind through unlikely proof strategies that humans might abandon. This approach can accelerate discovery, allowing you to focus on formulating novel questions and developing entirely new techniques, rather than being replaced by AI in the near term.

Key insights

AI disproved an 80-year-old math conjecture by combining existing techniques, showcasing its ability to resolve major open problems.

Principles

• AI excels at cross-domain knowledge application.
• AI tolerates tedious, low-probability proof attempts.
• Human-AI collaboration extends initial AI findings.

Method

The AI constructed a high-dimensional grid using algebraic integers, then projected this complex structure into two dimensions to demonstrate a point arrangement yielding more unit-distance pairs than traditional grids.

In practice

• Apply AI to problems needing broad knowledge.
• Utilize AI for exhaustive proof strategy exploration.
• Consider framing math problems for AI optimization.

Topics

• Erdős Unit Distance Conjecture
• Discrete Geometry
• AI in Mathematics
• Automated Theorem Proving
• Algebraic Number Theory
• Human-AI Collaboration

Code references

• teorth/erdosproblems

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

Related on AIssential

• OpenAI makes breakthrough on 80-year-old maths problem — OpenAI's AI disproved an 80-year-old math conjecture, showing AI can find novel solutions in complex problems.
• 😸 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.
• 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 Says Its AI Just Disproved a Famous 80-Year-Old Math Problem – And This Time, Mathematicians Agree — A general-purpose AI autonomously disproved an 80-year-old math conjecture by making a cross-domain mathematical leap.
• The AI Revolution in Math Has Arrived — AI is rapidly transforming mathematical research by accelerating discovery and proof generation, despite inherent model errors.
• An AI solution to an 80-year-old problem has shocked mathematicians — AI can autonomously solve long-standing mathematical conjectures by exploring vast solution spaces and combining existing knowledge.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.