Gemini Deep Think Earns Gold-Medal at the 2025 International Mathematical Olympiad
Google DeepMind's advanced Gemini Deep Think earned a gold-medal level score at the 2025 International Mathematical Olympiad (IMO) by solving five out of six problems, scoring 35/42 points. It is the first AI to receive such formal recognition.
Notably, OpenAI also announced a 35/42 score but did so without formal IMO certification, Gemini’s performance was officially verified.
The International Mathematical Olympiad (IMO) is “World Championship” of mathematics competition for High School students and is held annually. About 8% of participants achieve a gold medal. Google’s "Deep Think" mode used parallel reasoning, reinforcement learning focused on multi-step proofs, curated mathematical solution corpora, and competition-style hints. Gemini operated end-to-end in natural language within the 4.5-hour exam window.
Tom’s Take: Google’s breakthrough demonstrates AI’s ability to tackle complex mathematical reasoning directly with natural language, creating a foundation for advanced scientific applications. From an educational perspective, Deep Think’s achievement demonstrates how AI can assume the role of a proactive thinking partner, one able to guide learners through layers of math understanding.
Related posts from Tom Daccord’s free newsletter:
Also visit:
