OpenAI said on September 8, 2026, that its models had solved one of mathematics’ most famous problems, and the claim landed like a shock through a field already watching AI models produce proofs of decades-old conjectures all summer. The announcement did more than add another result to the record. It forced a fight over who gets credit, what counts as proof and whether human mathematicians are still the ones defining the work.
That is why the reaction has been so intense. Scott Aaronson wrote that he did not expect he would ever again prove a theorem because he was actually needed to prove it, and he added that human mathematicians were now forevermore dethroned as the main theorem-proving entities on planet earth. For readers looking up OpenAI Math today, the question is not only what was solved. It is whether the field has crossed a line that can be crossed only once.
Two days after the announcement, Ken Ono brought that anxiety into a lecture hall at the University of California, Berkeley. Ono had taken a leave of absence from the University of Virginia to work at Axiom Math, and he told about 150 students, postdocs and professors that they might be graduating into a profession that might not even exist, or that would be very different than they expected. The talk was scheduled for 50 minutes including questions. It lasted more than two hours.
The room did not respond like a crowd hearing a triumph. Students reacted with anger and frustration as Ono tried to explain the moment and what it might mean for mathematics. One student asked how Ono and his start-up would take responsibility in light of the way AI companies are treating mathematics. Another asked, in effect, what he was doing and what his very best was. Ono said he was not entirely sure what the takeaway was supposed to be, and he told the audience they needed to brace. The line that cut deepest came when he described the shameful way that AI companies are treating mathematics.
What followed was not just an argument over one result. It was a conversation about what mathematics really is and why mathematicians do it. The writer said many people in New York did not see what the problem was and thought amazing discoveries were around the corner, but the Berkeley exchange showed how unsettled the field remains. After the talk, students lingered to vent and console one another, while the larger dispute over authorship and standards stayed open.
That is the unresolved part of OpenAI Math: not the fact of a proof, but the terms on which the proof will be accepted, credited and absorbed into the profession. The announcement may have arrived as a breakthrough, but it also left mathematicians having to explain the value of their work in a world where machines are doing more of it.
