Archive · 23 Jul 2025
AI Models Win Gold at Math Olympiad, Advancing AI Capabilities
Imagine smart computer programs, like the ones that write stories, but for math problems!
This year, two of these programs from OpenAI and Google DeepMind competed in the International Math Olympiad, a super-tough math contest for high school students.
They did so well that they earned gold medals, just like the best human math students.
The programs worked by solving hard math problems step-by-step.
Some human math whizzes said these AI programs are good, but they still think people are better at coming up with brand new ideas for solving problems.
AI models from OpenAI and Google DeepMind achieved gold medal scores at the International Math Olympiad (IMO) 2025.
These AI models solved five out of six problems in the same conditions as human participants.
This is the first time AI models have achieved this level of success at the IMO.
Indian mathletes also won medals, with Kanav Talwar outperforming the AI models.
Researchers believe AI advancements could assist with solving complex problems in cryptography and space exploration.
- Who
- AI models from OpenAI and Google DeepMind competed against each other and human mathletes.
- What
- AI models from OpenAI and Google DeepMind achieved gold medal scores at the International Math Olympiad (IMO) 2025.
- When
- In 2025, at the International Math Olympiad.
- Where
- The International Math Olympiad (location not specified in articles).
- Why
- To showcase progress in AI and its problem-solving abilities.
AI Advancements
Human Capabilities
AI vs Human Problem Solving
AI Advancements
AI models have made significant progress in mathematical reasoning.
Human Capabilities
Human creativity and emotion in problem-solving are superior.
- Event
- International Math Olympiad (IMO) 2025
- AI Gold Medalists
- OpenAI, Google DeepMind
- Problems Solved
- 5 out of 6
- Medal Cut-off Score
- 35 out of 42 points
- Topics Covered
- Algebra, combinatorics, geometry, number theory