A Twitter post by polynoamial claims an internal OpenAI model called Astra has solved 10 major open problems in mathematics and computer science. The claim appeared on Hacker News, garnering 23 points and 5 comments. The original tweet links to a thread detailing the achievement, though no peer-reviewed publication or official OpenAI announcement has been referenced.

Commenters on Hacker News expressed skepticism, requesting verification and noting the lack of public benchmarks or paper citations. If true, this would represent a significant leap in AI reasoning capabilities, potentially impacting fields like cryptography, algorithm design, and theoretical physics. However, without reproducible evidence or formal disclosure, the claim remains unverified.

The discussion highlights ongoing interest in AI's ability to tackle long-standing scientific challenges.