On October 7 (Beijing Time), OpenAI declared that its state-of-the-art internal AI model has achieved multiple breakthroughs in mathematical research, subsequently releasing 722 mathematical manuscripts on Github. These manuscripts are neatly organized into 372 result families, spanning 17 distinct mathematical disciplines. Notably, some of these findings have undergone formal verification through Lean. The majority of these results were generated by OpenAI's unpublished, proprietary internal models, with the average computational investment for a single result comparable to that of ChatGPT Pro operating continuously for three hours. During the evaluation phase, the internal model was challenged with 4,000 mathematical queries. The academic community has greeted this news with a blend of awe and skepticism; while some scholars marvel at AI's prowess in mathematical exploration and advocate for embracing AI's integration, others highlight the relatively low proportion of high-quality, verified outcomes and counsel prudence. Furthermore, some academics have delved into the motivations driving AI companies' relentless pursuit of mathematical breakthroughs, voicing concerns over the potential repercussions for researchers in the field of pure mathematics.