An advanced version of Gemini with Deep Think achieved gold-medal performance at the International Mathematical Olympiad by perfectly solving five out of six problems, scoring a total of 35 points. This marks a significant improvement over the previous year, as the model now operates end-to-end in natural language, producing rigorous mathematical proofs within the competition's time limit. Google DeepMind aims to further enhance AI's capabilities in mathematics through ongoing collaborations and research advancements.