OpenAI's Math AI Has Now Solved 100+ Open Problems

OpenAI announced it is working with an independent Advisory Group on Mathematics and Artificial Intelligence, hosted at the Institute for Advanced Study, after disclosing that the internal model behind its disputed Navier-Stokes result has since resolved more than 100 long-standing open problems across most areas of mathematics, according to OpenAI's own announcement of the new board.

OpenAI's disclosure here is genuinely striking in scope. The company said it began training a new internal model on August 28, and beyond resolving the Navier-Stokes Millennium Prize problem we covered in detail at the time, that same model has now resolved more than 100 long-standing open problems spanning most areas of mathematics. OpenAI was candid that "the pace of its progress in mathematics has surprised the mathematicians within OpenAI," prompting internal discussions about how best to inform the broader mathematical community and help the field prepare and adapt.

This advisory group didn't emerge in isolation. It follows directly from an open letter titled "A Severe Misalignment of AI in Mathematics," published September 11 and signed by 25 Fields Medalists, mathematics' equivalent of a Nobel Prize, according to explainx.ai's reporting on the declaration. The letter, while not naming OpenAI or Anthropic directly, landed just three days after OpenAI's original Navier-Stokes announcement and the credit dispute it triggered. The signatories argued that AI companies' push to solve mathematical problems as a benchmark is "detrimental to the science of mathematics," according to Leiter Reports' coverage of the letter's core argument, warning that solutions announced in a rush leave no time for proper writeup, raise "severe attribution and plagiarism questions," and risk severing the human transmission chain that has traditionally carried mathematical understanding forward, not just final answers, between generations of mathematicians.

OpenAI was specific about the group's mandate, describing it as a genuine bridge to the mathematical community and broader public, giving mathematicians a real voice in the process. The group will advise on how to review and communicate emerging results, help assess their significance, coordinate dissemination, and weigh in on academic and professional research standards, alongside advising on how OpenAI's tools can support mathematical research and learning. Crucially, OpenAI stated directly that "the group will not be responsible for advising us on how to pace our internal progress on mathematics," a genuinely important boundary that separates this initiative from the broader industry pacing debate we've tracked closely this month, including our earlier coverage of the antitrust lawsuit alleging Anthropic, OpenAI, SpaceXAI, and Google illegally coordinated an AI slowdown.

The advisory group's initial nine members include Timothy Gowers of Collège de France and Cambridge, Edward Witten of the Institute for Advanced Study, and Ravi Vakil of Stanford, whose earlier comments on the original Navier-Stokes controversy we covered directly, alongside François Charles, Camillo De Lellis, Martin Hairer, Nikhil Srivastava, Ulrike Tillmann, and Melanie Matchett Wood. OpenAI emphasized the board's genuine independence directly in its announcement: "The group will operate independently from OpenAI. The group will have the freedom to offer advice we have not requested, comment on OpenAI's impact on mathematics, and make its advice public."

This development is worth understanding for any business in research-intensive fields evaluating how AI labs are responding to expert pushback on rapid, benchmark-driven capability claims, since OpenAI's decision to fund an unpaid, genuinely independent advisory board, one explicitly barred from influencing the pace of internal research, represents a notable governance template distinct from purely internal safety review processes.

For businesses in academic publishing, research software, or scientific computing, the disclosure that a single AI model has now resolved more than 100 open mathematical problems is worth watching closely as a signal of how quickly AI capability in formal, verifiable domains is genuinely accelerating, independent of the ongoing attribution and process disputes this specific announcement has triggered.

How many open math problems has OpenAI's model actually solved?
OpenAI disclosed that the internal model behind its Navier-Stokes result has resolved more than 100 long-standing open problems across most areas of mathematics since training began on August 28, 2026.

Why did mathematicians criticize OpenAI's approach to solving math problems?
Twenty-five Fields Medalists signed an open letter arguing that AI companies' rush to solve open problems as a capability benchmark bypasses the careful writeup, attribution, and community dissemination process central to how mathematical knowledge has traditionally advanced.

Will this new advisory group control how fast OpenAI develops its math AI?
No. OpenAI explicitly stated the group will not be responsible for advising on how to pace its internal progress in mathematics, limiting its mandate instead to reviewing and communicating results and advising on academic standards.

OpenAI announced an independent Advisory Group on Mathematics and Artificial Intelligence, hosted at the Institute for Advanced Study, after disclosing its internal model has resolved more than 100 long-standing open mathematical problems since training began August 28, a pace the company said surprised its own mathematicians. The board follows an open letter signed by 25 Fields Medalists warning that AI companies' rush to solve open problems as a benchmark bypasses proper attribution and the human transmission of mathematical understanding, landing just days after OpenAI's disputed Navier-Stokes announcement. The unpaid, genuinely independent group can publish criticism of OpenAI unprompted, though its mandate explicitly excludes advising on how fast OpenAI should pace its internal mathematical research.