Humans Outperform AI at International Math Olympiad, But Artificial Intelligence Closes the Gap

For the first time, artificial intelligence models from Google and OpenAI earned gold-level scores at the International Mathematical Olympiad, a renowned global mathematics contest for students under 20. While five human competitors achieved perfect marks, the AI models displayed rapid progress, with Google DeepMind’s Gemini chatbot solving five out of six problems and meeting the competition’s rigorous time constraints—an improvement over last year’s slower performance. Observers lauded the AI’s clear and precise mathematical proofs, raising questions about the accelerating development of AI and its potential to reach, or even surpass, human-level cognitive abilities. Industry leaders predict that exponential advances and mounting investment in the field could yield artificial general intelligence within the next decade.

Read more