OpenAI and Gemini (DeepMind) both gained enough score to win gold medal on the 2025 International Mathematics Olympiad. IMO 2025 Problem 6 was a difficult challenge that they both failed on, but this advancement in AI shows how IMO benchmark is slowly attainable even by commercial SOTA models that we might face today. The IMO officials also scored Gemini's answers from two models that Gemini brought forward while OpenAI ran it privately with independent grading from previous participants. #ai #openai #gemini #internationalmathematicalolympiad