Google Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed

Google DeepMind has published a study in Science describing AlphaProof Nexus, an AI system it says solved dozens of open mathematics problems on its own. According to the report, the system cracked nine of 353 open Erdős problems, two of which had gone unsolved for 56 years. The team shared those solutions online ahead of the paper, and the work went through peer review before publication.
Beyond the Erdős problems, AlphaProof Nexus proved 44 conjectures drawn from the OEIS integer sequence database, solved an open question in algebraic geometry, and improved a known bound in min-max optimization, a technique used to train some AI models. Google DeepMind reports that the system solved the nine Erdős problems even without calling on AlphaProof, its specialized proof tool.
The system is built around Gemini 3.1 Pro, which drafts and revises proofs in Lean, a formal language for mathematics that a machine can check. Lean verifies each draft after every edit, and the error messages it returns guide the next attempt. For harder sub-problems, AlphaProof Nexus can bring in AlphaProof, the specialized proof tool.
The timing is notable. OpenAI recently posted hundreds of AI-generated proofs on GitHub rather than submitting them to journals, prompting the Association for Human Mathematics, led by Terence Tao, to call for a boycott. DeepMind instead put its results through peer review.
Why it matters: Formal verification via Lean means these proofs are machine-checkable rather than relying on human reviewers to catch every gap, which is a meaningful difference from posting unverified proofs online. For developers and product teams, the peer-reviewed publication and the boycott controversy together signal that verification standards are becoming a dividing line in AI math research. Whether that pressure changes how labs release results is, in my reading, still an open question.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
Google Deepmind's AlphaProof Nexus solved dozens of open math problems, including nine Erdős problems two of which were unsolved for 56 years, per a study in Science. The publication comes as OpenAI faces a boycott for dumping AI proofs on GitHub without peer review. Deepmind had its results reviewed before publishing. The article Google Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed appeared first on The Decoder .