NYU mathematician Tristan Buckmaster says OpenAI privately informed him on Sept. 6 that an internal model produced a 100 page proof for the Navier-Stokes Millennium Prize Problem. The disputed breakthrough involves a $1 million prize and sparked a conflict over scientific credit, with Buckmaster alleging that OpenAI researchers misrepresented the level of human effort and compute required. Buckmaster also raised questions about whether OpenAI trained the model on private Codex sessions used by him and collaborator Levent Alpöge.

OpenAI researcher Sebastien Bubeck described the allegations against him as "false and inflammatory" after claims surfaced that OpenAI attempted to remove Alpöge from the authorship list due to his employment at Anthropic. Buckmaster reports that OpenAI staff asked "Why would you ruin your career?" when he threatened to reveal the private talks. OpenAI is scheduled to release a statement regarding its progress on the Navier-Stokes problem on Sept. 9.