In a remarkable milestone for artificial intelligence, two advanced AI models — GPT-5 and Gemini 2.5 Pro — have achieved gold medal performance at the International Olympiad of Astronomy and Astrophysics (IOAA), according to research published in October 2025. This achievement highlights how far AI has progressed — from generating language to solving complex scientific problems once thought to require human intuition and expertise. AI Excels at Complex Physics Problems Researchers tested several large language models using past IOAA theoretical exams . These exams are known for their difficulty, demanding deep understanding, logical derivation, and data interpretation. Among all tested models, GPT-5 and Gemini 2.5 Pro stood out with exceptional scores: GPT-5 scored 85.6 percent, Gemini 2.5 Pro scored 84.2 percent. Both results surpassed the gold medal threshold, placing the models in the top tier of all participants — a level equivalent to human gold medalists. However, the stud...