Google has launched the Gemini 2.5 Deep Think AI model, achieving gold-medal standards at the International Math Olympiad (IMO) held in Australia. This milestone marks the first instance where AI, including Google’s Gemini and OpenAI, surpassed the gold score for high school levels. Despite this success, the model currently requires hours to resolve complex math problems and has only attained bronze-level performance based on internal projections for the 2025 IMO. Google is piloting the Deep Think model with select mathematicians to gather feedback for enhancement. Utilizing parallel thinking techniques, the Gemini model can generate and evaluate multiple solutions simultaneously. This innovative approach not only augments creative thinking but also aids researchers in navigating complex scientific literature. Additionally, the model achieves state-of-the-art performance in various benchmarks, illustrating its potential to significantly impact the field of mathematics and coding. Sundar Pichai emphasizes that AI has revolutionized access to extensive research for global communities.
Source link
Share
Read more