OpenAI has withdrawn three mathematics papers from its public record, according to the change history maintained in the company's math research repository on GitHub.
The available source identifies the withdrawals but does not provide details on the papers' subject matter, the timing of their removal or OpenAI's reasons for taking them down.
Paper withdrawals can reflect issues ranging from errors in methodology to incomplete results or publication decisions. Without further disclosure, however, the record alone does not establish why OpenAI withdrew the work.
The development highlights the scrutiny faced by AI companies when they publish technical research, particularly in mathematics and reasoning — areas increasingly central to claims about advanced model capabilities.