Archive — US — techcrunch
OpenAI and Google outdo the mathletes, but not each other
OpenAI and Google DeepMind achieved gold medals at the 2025 International Math Olympiad, showcasing their advanced AI systems. While both companies scored higher than most high school students, conflicts arose over announcement timing and evaluation methods, highlighting intensifying competition and differing approaches in the AI landscape.
Key points
- OpenAI and Google DeepMind both claimed gold medals in the 2025 IMO.
- The achievements reflect rapid advancements in AI reasoning capabilities.
- OpenAI's method involved informal systems with no human-machine translations, improving from their previous year's performance.
- Google criticized OpenAI for premature announcement and lack of official approval for their testing system.
- The competition reveals a close race between AI companies for public perception and leading talents in the field.
Why it matters
- The situation mirrors historical competitions in AI development, such as the AlphaGo vs. Lee Sedol match, indicating how public perception affects talent acquisition.
- In 2025, fierce competition in AI continues, where advancements in natural language processing and reasoning are crucial.
- Similar to the 2025 emphasis on AI ethics, the controversy underscores the importance of verification and transparency in AI model performance.
The gold medal performances by OpenAI and Google illustrate significant progress in AI technology and emphasize the escalating competition for dominance in the AI sector, as both companies navigate challenges in credibility and public perception.