Humans outperformed generative AI models developed by Google and OpenAI at the 2025 International Mathematical Olympiad (IMO), even though machines achieved gold-level scores for the first time in the competition's history.
The IMO, held in Queensland, Australia, featured six challenging problems to be solved within 4.5 hours for participants under the age of 20. While Google's Gemini and OpenAI's reasoning model each scored 35 out of a total of 42 points, enough for a gold medal, neither achieved a perfect score.
In contrast, five human competitors earned full points.
Google announced on Monday, July 21, that its advanced chatbot Gemini successfully solved five of the six problems, marking a significant improvement over the previous year, when it only achieved a silver-level score at the IMO in Bath, UK.
“We can confirm that Google DeepMind has reached the much-desired milestone, achieving 35 points out of a total of 42, a gold medal score,” said IMO President Gregor Dolinar.
“Their solutions were astonishing in many ways. IMO graders found them clear, precise, and most of them easy to follow.” OpenAI, the developers of ChatGPT, also reported a gold-level performance, with its experimental reasoning model scoring 35 points.
“We evaluated our models on the IMO 2025 problems according to the same rules as the human competitors,” said OpenAI researcher Alexander Wei, noting that each submission was independently scored by three past IMO medalists.
Google noted a significant reduction in computational time this year, with its model solving the problems within the competition’s allotted time. In contrast, last year’s attempt took two to three days of computation. The IMO confirmed that companies had privately tested closed-source AI models on the same problems attempted by 641 students from 112 countries.
"It's very exciting to see advances in the mathematical capabilities of AI models," Dolinar said.However, he warned that competition organizers could not verify the scale of computing resources used by AI or whether human assistance had been involved.
Noam Brown, a senior OpenAI researcher who worked on the IMO model, said IMO contacted OpenAI a few months ago to participate in a formal math competition, but the creator of ChatGPT declined because he was working on natural language systems that he found more interesting to pursue. Brown says OpenAI didn't know IMO was conducting an informal test with Google.
OpenAI is expected to release GPT-5 in the coming month.