"AI Mathematical Olympiad 2 Concludes: NemoSkills Solves 34 Out of 50 Challenging Problems, Surpassing AIMO1 Record"

The second Artificial Intelligence Mathematical Olympiad (AIMO2) competition has recently concluded. I am a member of the advisory committee for this competition, which took place on Kaggle. You can explore the details of the event at this link: kaggle.com/competitions/ai-mathematical-olympiad-progress-prize-2/overview. The AIMO2 featured 50 test problems designed to be more challenging to brute force methods compared to the first iteration, AIMO1. However, these problems remained in the same "numerical response" format, requiring contestants to provide specific numerical answers. Participants were mandated to use open-source models and were given a fixed amount of computational resources to ensure a level playing field. While final certification is still pending, the presumptive winner is NemoSkills, who successfully solved 34 out of the 50 problems. For context, the best performer in AIMO1 managed to solve 29 out of 50 problems, marking a significant improvement in AI capabilities over a relatively short period. We are now in the process of evaluating the performance of more advanced models, as well as those allowed greater computational resources. Initial results suggest substantial improvements, and we will be sharing more details in upcoming announcements. Stay tuned for further updates on the progress and findings of this innovative competition.