Archive · 23 Jul 2025
Humans Outperform AI at Math Olympiad Despite Record Scores
Imagine a super-hard math test for the smartest kids in the world, called the IMO.
This year, super-smart computers (AI) from Google and OpenAI tried to take the test too.
They did pretty well and got gold medals!
But guess what?
Five human kids solved every problem perfectly, beating the AI.
This shows that even though AI is getting better, human brains are still the best at solving really tricky problems and thinking in creative ways.
Humans outperformed AI models from Google and OpenAI at the International Mathematical Olympiad (IMO).
Five human mathematicians achieved perfect scores, while the AI models scored gold-level but not perfect.
Google's Gemini chatbot solved five out of six problems, achieving a gold medal score.
OpenAI's reasoning model also scored a gold-level, with 35 points.
The IMO highlights human intuition's edge in complex problem-solving.
- Who
- Human mathematicians and AI models from Google and OpenAI.
- What
- Humans outperformed AI in the International Mathematical Olympiad.
- When
- The International Mathematical Olympiad (IMO) was held this month.
- Where
- Queensland, Australia
- Why
- To assess the capabilities of AI in complex problem-solving compared to human abilities.
AI Progress
Human Excellence
AI Capabilities vs. Human Capability
AI Progress
AI models are making great strides in complex problem-solving.
Human Excellence
AI has yet to surpass humans in this area
- Event
- International Mathematical Olympiad (IMO)
- Location
- Queensland, Australia
- Human Perfect Scores
- 5
- Google Gemini Score
- 35/42 points
- OpenAI Model Score
- 35 points
- Number of countries competing
- 112