Archive · 28 Jul 2025
High School Students Outperform Advanced AI Models at Math Olympiad
Some very smart computers (AI) were tested on a tough math test with high school students.
The computers, made by Google and OpenAI, did great and won gold medals!
However, the brightest students beat the AI.
These AI can solve math problems really well and even understand the problems like humans do.
The AI had trouble with the hardest problem that required clever thinking, which is something the smart students excelled at.
This shows that while AI is improving, humans are still needed to be at the cutting edge.
AI models from Google DeepMind and OpenAI earned gold-medal scores at the International Mathematical Olympiad (IMO).
Despite the AI achievements, 26 high school students scored higher than the AI systems.
Qiao (Tiger) Zhang and Alexander Wang, both U.S. team members, were among the top-scoring students.
DeepMind's AI solved problems in natural language, showing significant progress.
The AI models struggled with Problem 6, which required human ingenuity.
- Who
- Google DeepMind, OpenAI, and high school students, including Qiao Zhang and Alexander Wang, participated.
- What
- AI models and high school students competed in the International Mathematical Olympiad (IMO).
- When
- The competition occurred recently, with results announced shortly after the event.
- Where
- The International Mathematical Olympiad (IMO) took place in Australia.
- Why
- The competition served as a benchmark for AI progress and a test of mathematical abilities.
Human Perspective
AI Perspective
Capabilities in complex problem solving
Human Perspective
Human creativity, intuition, and ingenuity are superior.
AI Perspective
AI can now complete the same tasks as humans.
- Competition
- International Mathematical Olympiad (IMO)
- AI Participants
- Google DeepMind, OpenAI
- Human Participants
- Elite high school students
- Gold Medalists (AI)
- Google DeepMind, OpenAI
- U.S. Team Stars
- Qiao (Tiger) Zhang, Alexander Wang
- AI's Problem
- Problem 6