The competition for AI supremacy has intensified as Google DeepMind and OpenAI claim their models can tackle the International Mathematical Olympiad (IMO), a prestigious intellectual challenge. Google’s Gemini Deep Think AI secured an officially certified gold medal at the 2025 competition, scoring 35 out of 42 points, while OpenAI also announced its model achieved the same score unofficially just days prior. This milestone demonstrates significant advancements in AI reasoning, establishing that general-purpose models can solve intricate mathematical problems at an elite human level.
Both companies emphasize their models’ capabilities extend beyond mathematics, with Google leveraging its new Deep Think architecture and OpenAI highlighting experimental general-purpose techniques. As these teams prepare for wider testing, the implications reach beyond mathematics, potentially revolutionizing discoveries across various scientific fields. This breakthrough brings us closer to achieving Artificial General Intelligence (AGI), marking a pivotal moment in the evolution of AI capabilities.
Source link
