OpenAI recently announced a groundbreaking achievement: its experimental AI reasoning model won a gold medal at the 2025 International Math Olympiad (IMO), one of the world’s most prestigious math competitions. Competing under standard conditions, the model solved five of six complex problems, scoring 35 out of 42 points, which aligns with the gold medal cutoff. This success demonstrates the potential for AI to excel beyond specific tasks, indicating significant advancements in general intelligence. Notably, OpenAI’s model generated natural language proofs independently, with scores verified by former IMO medalists. This milestone comes amid rising competition, with Google DeepMind’s AlphaGeometry 2 also claiming to achieve gold-level performance. Despite the impressive results, some critics, like NYU’s Gary Marcus, question the lack of independent verification. OpenAI’s CEO, Sam Altman, emphasized that these advancements in AI math capabilities are integral to their pursuit of general intelligence, though the model won’t be publicly available for several months.
Source link
OpenAI’s Next Major Model: A Game Changer for Winning International Math Olympiad Gold? | Technology News
Share
Read more