\n\n
A landmark batch of 722 research papers generated by an unreleased OpenAI model has been posted to GitHub, covering roughly 20 mathematical subfields. Among the highlights, the system produced a proof of the quasi‑Riemann hypothesis, an important step toward the famed Riemann hypothesis that underpins many areas of number theory. The effort also yielded over 80 theoretical‑computer‑science papers, new definitions of limits for matrix‑multiplication speed‑ups, a novel integer‑multiplication algorithm, and more than a dozen proofs concerning partial differential equations, including clarifications of Navier–Stokes questions. These advances demonstrate AI’s growing capacity to contribute original insights across pure and applied mathematics, and they set the stage for further verification, formal Lean proofs, and collaborative research funded by OpenAI. [[1]]