OpenAI's 722-paper math release tests research credibility

OpenAI has publicized 722 mathematical manuscripts and Lean formalizations, with Altman framing the work as evidence of a new era of discovery. Independent commentary emphasizes that some claims remain incompletely checked or difficult to reproduce, making validation, disclosure of prompts and reasoning, and scientific credibility the central open questions.