OpenAI says its AI system advances 10 major math problems

OpenAI said its AI system has advanced 10 major mathematics problems, according to a report from The American Bazaar. The claim, relayed from OpenAI's own statements, ranks among the company's most high-profile capability announcements in recent months.

The problems in question have reportedly resisted solution for years, with several remaining open for more than a decade. OpenAI framed the results as evidence that its AI system can now make genuine headway in mathematical research rather than merely score well on benchmarks.

So far the details are thin: the full solutions and proof documents have not been made public, and the American Bazaar article mostly relays OpenAI's characterization of the results. Independent verification will have to wait for the underlying materials.

This is not the first time OpenAI has claimed a mathematics breakthrough. Over the past year the company has repeatedly reported progress in mathematical reasoning, and mathematicians have grown more cautious about such announcements, calling for rigorous scrutiny of any AI-generated proof.

If the results hold up, the implications extend well beyond mathematics, and AI could shift from a problem-solving tool into a research partner capable of independently discovering new theorems, with knock-on effects for theoretical computer science and other fields that rely on formal proofs.

The immediate question is verification. Mathematicians will need to check whether each result is correct, whether the proofs are rigorous, and whether the system's approach can be reproduced, work that often takes longer than the announcement itself.

The next thing to watch is whether OpenAI publishes the detailed proofs and papers, and when the AI system becomes publicly available. Either way, the statement has pushed the question of AI participation in frontier mathematics back into the spotlight.