AI Models From Google, OpenAI Win Gold At International Math Olympiad

Google and OpenAI have announced that their artificial intelligence (AI) models have achieved gold medal-level at the International Mathematical Olympiad (IMO), a prestigious competition for high school students.

The AI systems from both companies successfully solved five out of six problems, using advanced “reasoning” models capable of processing complex mathematical concepts through natural language.

While OpenAI did not officially participate in the competition, Google DeepMind worked with the IMO and was graded by the official committee. OpenAI’s results were graded by three external IMO medalists, according to a report by Reuters.

This breakthrough suggests that AI may be less than a year away from assisting mathematicians in solving unsolved research problems, according to Junehyuk Jung, a mathematics professor at Brown University and a visiting researcher at Google DeepMind.

“I think the moment we can solve hard reasoning problems in natural language will enable the potential for collaboration between AI and mathematicians,” Jung told Reuters.

OpenAI researcher Noam Brown explained that their breakthrough was enabled by a new experimental model built around a “test-time computer.” The model was allowed to “think” for longer periods and leveraged parallel computing to explore multiple lines of reasoning simultaneously.