OpenAI has informed researchers it intends to release hundreds of solutions to unsolved mathematical problems on GitHub this week, according to a report published by Wired on October 6, 2026. The planned drop would mark the latest wave in tens of thousands of AI-generated mathematical solutions released this year as frontier models grow more capable. Leading mathematicians say the company has disregarded explicit advice they gave during a private meeting in August, when OpenAI convened around 40 researchers to discuss how to handle the situation if AI surpasses human abilities in mathematics.

During that August gathering, OpenAI suggested its advanced models had cracked hundreds of longstanding math problems, people who attended told Wired. The group urged the firm not to simply post them in a blog or tweet, as it had done with 10 problems earlier that month, recalls Northwestern University mathematician Bryna Kra. Instead, attendees emphasized OpenAI should publish papers explaining the work so mathematicians could absorb, digest, and apply the results. Company representatives assured attendees it wouldn't release the solutions all at once—though spokesperson Lindsay McCallum says the company is "not aware of" making such an assurance—and asked for guidance on how to publish the findings. Attendees reacted with "a mixture of excitement and dread," but viewed the session as a promising first step. "Apparently, that input was ignored," Kra told Wired.

The controversy deepened in September when OpenAI deployed thousands of agents to solve a legendary million-dollar Millennium Prize problem after hearing "rumors" that others were closing in on solutions. Tristan Buckmaster, a New York University mathematician and professor, accused the company of front-running work he had done toward solving an element of the problem in a personal collaboration with Anthropic employee Levent Alpöge. The pair hadn't published their work but had been using OpenAI's tools to help them. According to meeting notes seen by Wired, OpenAI researcher Sébastien Bubeck appeared to suggest during negotiations that Alpöge should be excluded from any paper because that would make things complicated—though Bubeck has publicly denied asking for Alpöge to not be listed as an author. Buckmaster said he refused. Visiting NYU math professor Nestor Guillen told Wired that "there's a perception of mobster behavior" from the AI companies among mathematicians, though McCallum says the company disagrees with that characterization.

Mathematicians say the core issue is that OpenAI and Anthropic continue releasing results through blog posts rather than scientific papers, making them harder to verify and often excluding prior work from other researchers. The report notes that OpenAI assembled an advisory group in mid-September to help determine how the company assesses and communicates new results, upholds academic and professional standards, and builds tools that support mathematical research and learning. But mathematicians say the company hasn't made much progress. "Math by tweet and math by press release to me is not the way to nurture the ecosystem that created the fertile ground that they have trained on," Kra said. While OpenAI has been directly encouraged to use new community tools created this year to help researchers both use new results and sort through what needs further review—including Hexagon, a repository for primarily AI-generated material, and Palomar, a registry of machine-verified mathematics—Kra notes "they haven't changed their behavior."

Researchers told Wired they've interpreted OpenAI's private briefings as warnings about what's to come, with the company describing its capabilities and breakthroughs to mathematicians the way "police might want to notify family before reporting a death in a car accident." Several employees within OpenAI believe their technology has rendered math dead anyway, according to people who have spoken with them, though McCallum says the company doesn't believe the future of mathematics is set and is working with the math community to navigate the future collaboratively. Mathematicians know they need to adapt, especially to equip younger generations, but they don't believe their field is dead. "It changes how we're going to operate, but I think it's a moment that we can think bigger," Kra said. She's happy to have a new, powerful tool at her disposal to solve problems—she just wants companies' disclosures to be better so that she can trust and build on the results. The battle over how AI labs share breakthroughs may ultimately determine whether researchers view these tools as collaborative partners or extractive competitors. What's clear is that the traditional academic norms that governed scientific progress for centuries can't be enforced when the institutions generating discoveries operate on Silicon Valley timelines rather than peer review calendars.