OpenAI and mathematicians dispute Navier-Stokes proof authorship after misconduct claims

September 9, 2026 at 03:27 PM2 min

This digest was compiled by AI from multiple sources — links to the originals are below.

Mathematician Tristan Buckmaster accused OpenAI of misconduct over its AI-generated proof of the Navier-Stokes equations, a Clay Millennium Problem with a $1 million prize. OpenAI CEO Sam Altman defended the team, saying they acted with integrity and generosity. The dispute centers on whether OpenAI used the researchers' drafts to train its models and whether the proof is independent.

Key Facts

The Navier-Stokes equations are one of the Clay Millennium Problems, with a $1 million prize for a correct solution.
OpenAI CEO Sam Altman said the team acted with integrity and generosity throughout and suggested Buckmaster and Alpöge should receive the award.
OpenAI employee Boaz Barak said the model started by proving a stronger claim than Buckmaster and Alpöge did.
OpenAI's official blog post states that while unlikely, it cannot rule out that de-identified data from the researchers' usage helped improve its models.
Altman said the effort began after rumors on the internet that Anthropic's models had solved a millennium problem.

The Dispute

Mathematician Tristan Buckmaster accused OpenAI of misconduct after the company released an AI-generated proof of the Navier-Stokes equations. Buckmaster claimed that after information about his research leaked, OpenAI pressured him and tried to remove his co-author Levent Alpöge from the paper because Alpöge works at Anthropic. Buckmaster also said OpenAI may have trained its models on drafts the two researchers uploaded to Codex. Buckmaster called the situation 'absolute academic malpractice'.

OpenAI's Response

OpenAI CEO Sam Altman backed the team on X, saying they had acted with integrity and generosity throughout. Altman said it had been suggested that Buckmaster and Alpöge should receive the award, but the team was met with unfounded accusations of plagiarism. OpenAI employee Boaz Barak pushed back on the idea that the model needed outside help, saying it started by proving a stronger claim than the researchers did. OpenAI's official blog post acknowledged that while unlikely, it cannot rule out that de-identified data from the researchers' usage helped improve its models.

Conflicting Accounts

Alpöge and Buckmaster dispute OpenAI's claim that its solution differs significantly from theirs, saying they had entered a similar approach into OpenAI's systems. Altman wrote that it was difficult to make the same offer to Alpöge because he was not willing to talk or coordinate with OpenAI. Alpöge said he would have liked to work with OpenAI and that the authorship question did not matter to him. Altman said the effort started because of rumors on the internet that Anthropic's models had solved a millennium problem.