OpenAI and Google AI Tie for Gold at 2025 International Math Olympiad
AI Rivals OpenAI and Google Achieve Gold at 2025 International Math Olympiad
Artificial intelligence has reached a new milestone in mathematical problem-solving. Both OpenAI and Google DeepMind recently announced that their AI systems achieved gold medal-level scores in the 2025 International Math Olympiad (IMO)—one of the world's most challenging math competitions for high school students.
AI Models Compete—and Excel—at Olympiad Level
This year, both companies entered "informal" AI systems to the IMO. Unlike previous formal systems that required humans to translate math problems into machine-readable formats, the latest models could directly process natural language questions and generate proof-based answers. The results? Both OpenAI and Google’s AI models outperformed most human contestants and surpassed Google’s own silver-medal-winning AI from last year.
- OpenAI and Google entered their models independently
- Both achieved gold medal scores, matching the top human competitors
- No human translation was needed; models processed natural language questions directly
A Tight Race in AI Research
The neck-and-neck results underscore an intense rivalry in AI development. As both companies vie for leadership—and for the top talent in the industry—achievements like these are more than just technical milestones. Many AI researchers come from math competition backgrounds, making benchmarks like the IMO especially meaningful within the AI community.
Debate Over Announcement Timing
Despite their similar achievements, the timing and manner of the announcements sparked debate. Google DeepMind CEO Demis Hassabis publicly criticized OpenAI for announcing its results before the IMO’s official verification and before students had received their awards. Google stated that it waited until after the formal grading and the IMO president’s approval to make its announcement, out of respect for the student competitors.
OpenAI, on the other hand, had their model evaluated by three former IMO medalists and reached out to the IMO organizers after receiving their gold-level results. However, Google asserts that only official IMO grading can confer true gold medal status.
What This Means for AI and Business
While the debate over protocol is notable, the bigger story is the rapid progress in AI reasoning and problem-solving. Being able to solve Olympiad-level problems in natural language demonstrates significant advances in AI’s ability to tackle complex, structured tasks—a capability with broad implications for businesses using AI for research, analytics, or automation.
The results show that AI is catching up with (and in some cases surpassing) the abilities of top human experts in highly specialized domains. For business owners, this evolution signals new opportunities to leverage AI for advanced problem-solving, optimization, and innovation.
Looking Ahead
With OpenAI expected to release GPT-5 soon and Google continuing to refine its Gemini models, the competition between these AI giants is poised to accelerate. As AI models continue to improve, expect even more breakthroughs—not just in math, but across many fields demanding high-level reasoning and creativity.
