AI Frontiers Digest

AI Breakthroughs in Mathematics: OpenAI, DeepMind, and xAI Solve Open Problems

AI Breakthroughs in Mathematics: OpenAI, DeepMind, and xAI Solve Open Problems

Key Questions

What math conjecture did OpenAI's reasoning model disprove?

OpenAI's reasoning model disproved the 80-year-old planar unit distance conjecture using higher-dimensional lattices. The result was validated by mathematicians Gowers and Bloom.

How many Erdős problems did DeepMind AlphaProof Nexus solve?

AlphaProof Nexus solved 9 Erdős problems through formal proof generation. This demonstrates AI's growing capability in solving long-standing open mathematical questions.

What open problem did Grok 4.5 solve?

Grok 4.5 solved the hypercontractivity problem on the 4-sphere, closing a long-standing question in functional analysis. It highlights xAI's contribution to theoretical mathematics alongside OpenAI and DeepMind.

OpenAI reasoning model disproves 80-year-old planar unit distance conjecture using higher-dimensional lattices; validated by Gowers/Bloom. DeepMind AlphaProof Nexus solves 9 Erdős problems via formal proof. Grok 4.5 (xAI) solves hypercontractivity on the 4-sphere, closing a long-standing open problem in functional analysis. A tutorial on proving theorems with Lean and ML at ICML 2026 highlights formal verification's growing role. AI is emerging as a co-discoverer in theoretical research.

Sources (2)
Updated Jul 10, 2026