AI Breakthrough: Google's Gemini Deep Think and OpenAI Models Excel at Global Math Olympiad

Artificial intelligence has scored a significant victory at the International Mathematical Olympiad as Google's Gemini Deep Think and models from OpenAI captured gold medals, a first in AI history. Competing in Queensland, Australia, both companies' AI successfully solved five out of six problems, showcasing a breakthrough in mathematical reasoning capabilities.

The achievement was described by Junehyuk Jung, a math professor visiting Google's DeepMind, as a potential game-changer for collaboration between AI and mathematicians to tackle unsolved research problems. OpenAI's new experimental model utilized extensive computing power and longer processing times to think through complex problems, although the exact cost of resources remains undisclosed.

This success is seen as a promising indicator that AI's reasoning abilities will extend beyond mathematics into other scientific realms, according to Google researchers. With AI systems operating in natural language, the event marked formal collaboration with the International Mathematical Olympiad's judging, as noted by various stakeholders including Google's DeepMind CEO Demis Hassabis.