Google’s Gemini AI Clinches Gold at the 2025 Mathematics Olympiad with Official IMO Accreditation

Google DeepMind’s Gemini Deep Think model achieved a historic gold medal at the 2025 International Mathematical Olympiad, outperforming its previous systems and showcasing its evolution into a mathematical collaborator. Under stringent time constraints, the AI successfully solved five of six problems, earning 35 out of 42 points—marking a significant improvement from last year’s silver medal score of 28 points. This accomplishment is noteworthy for its official validation, having undergone rigorous evaluations identical to those of human competitors. Key advancements in Gemini’s architecture, including Parallel Thinking and Reinforcement Learning, enable it to solve problems using natural language reasoning, resembling human mathematical thought processes. This breakthrough has sparked debate regarding AI’s role in mathematics, with potential implications for various commercial applications. Investors are closely watching as the 25% jump in reasoning capability could open new markets in AI-powered verification services and educational technology, redefining the landscape of mathematical research and application.

Source link