Google DeepMind’s advanced Gemini Deep Think AI has reached a milestone by solving five out of six problems at the 2025 International Mathematical Olympiad, achieving gold-medal status. Unlike previous models, Gemini operated entirely in natural language and within official time limits, highlighting substantial progress in AI mathematical reasoning and proof-generation capabilities. This significant milestone was reached by solving five out of six problems in the competition, showcasing substantial advancements in AI mathematical reasoning and proof generation capabilities. The advanced Gemini model scored 35 out of a possible 42 points, with solutions praised for their clarity and precision, signalling an improvement over previous entries. After welcoming listeners to the episode, Cochrane shares a sponsored message from GoDaddy, highlighting exclusive deals for listeners.