Google Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed
decoder ·
A new study published in Science reveals that Google DeepMind's AlphaProof Nexus successfully solved dozens of open mathematical problems. The AI system cracked nine out of 353 open Erdős problems, including two that had remained unsolved for 56 years. Additionally, it proved 44 conjectures from the OEIS integer sequence database and resolved an open question in algebraic geometry. The model is built around Gemini 3.1 Pro and utilizes the formal language Lean to draft and verify proofs iteratively. For complex sub-problems, it can leverage AlphaProof, a specialized proof tool, though it achieved significant results even without it. This rigorous approach contrasts with recent controversies where other AI firms posted proofs directly to GitHub without peer review. In response to calls for boycotts by mathematicians, DeepMind ensured its work underwent traditional academic scrutiny.
decoder