Three OpenAI math papers withdrawn after one shared error
Retraction Watch reports that OpenAI withdrew three math papers after a sign error affected all three. It also notes 14 more revisions after a large release of manuscripts. Researchers welcomed the quick corrections but raised questions about how the work was released and checked. These cases show why changes need a clear record. They do not tell us how many other papers are wrong.
What the source reports
Retraction Watch reports that OpenAI withdrew three math papers after a sign error appeared in work they shared. The report also notes 14 more revisions after a large release of manuscripts. The researchers it interviewed welcomed quick corrections. They also questioned how such a large batch of work was released and checked. A correction matters most when readers can see which papers depend on a changed result. But three withdrawals are not a sample of all the remaining work. They show a need for a clear record of errors and links between papers, not a measured error rate for the whole release.
Original source
- Title
- OpenAI withdraws three preprints a day after releasing 722 manuscripts on unsolved math problems
- Author
- Retraction Watch
- Publication
- Retraction Watch
- Date
- Thursday, October 8, 2026