Story
September 23, 2026
OpenAI’s Math Claims Put Its New Advisory Panel to the Test
OpenAI says an internal model solved the Navier–Stokes problem and more than 100 other open questions. Its new independent mathematics panel is meant to build trust, but researchers want proof that elite advisers will have real influence.
OpenAI says the turning point began on August 28, when it started training a new internal model. The company now claims that model has resolved the Navier–Stokes Millennium Prize problem as well as more than 100 long-standing open problems across most areas of mathematics—progress it says surprised its own mathematicians.1
Those claims created a problem beyond the proofs themselves: how to disclose potentially consequential results without treating the discipline as a benchmark laboratory. OpenAI acknowledged criticism from mathematicians who warned of the negative effects of using unsolved problems to measure AI systems, saying the backlash underscored the need for serious engagement with the field.1
Its answer is the Advisory Group on Mathematics and Artificial Intelligence, a nine-member body hosted at the Institute for Advanced Study. OpenAI says the panel can review emerging results, advise on their significance and dissemination, comment publicly on the company’s impact, and offer unsolicited advice. Members will not be paid by OpenAI and may change the group’s membership themselves. But the company also draws a firm boundary: the group will not advise on the pace of its internal mathematical progress. “Its value depends on its members being able to exercise their own judgement and challenge ours,” OpenAI said.1
For many human mathematicians, that independence is promising but unproven. The panel’s sudden launch followed what The Verge described as a reputational crisis over OpenAI’s earlier mathematical announcements. Researchers told the publication it was “a good first step,” while raising basic questions about its authority, its representation of the wider community, and whether OpenAI will act on its advice.2
The tension is sharp: OpenAI is offering access to some of mathematics’ most prominent figures, while critics are asking whether prestige can substitute for accountability. The panel may help shape how discoveries are communicated; whether it can shape the company’s choices remains the harder test.