uncommented | 14:05 GMT+3 – 14:25 GMT
The 2025 International Mathematical Olympiad (IMO), the world’s most prestigious math competition, brought together the best mathematical minds to test the limits of artificial intelligence (AI). This year, AI models by leading companies such as Google’s DeepMind and OpenAI’s ChatGPT demonstrated remarkable capabilities. Specifically, Google’s DeepMind’s Gemini Deep Think and OpenAI’s OpenAI’s Gemini Deep Think each solved five out of six problems in the IMO, earning gold medals for both models. Meanwhile, OpenAI’s AI model also achieved a similar score, while DeepMind’s Gemini Deep Think secured a second gold medal. The IMO itself was perceived as an monumental challenge, with only a relatively small percentage of participants awarded gold medals, specifically 10 out of 630 competitors. This test highlights the extreme difficulty IMO has selected, indicating a strict selection process with a twin focus on human participants and the achievements of those who are modeled after Olympiad winners.
The IMO is primarily hosted by elite high schools for students aged under 20. This year saw an influx of participants, with more than 100 countries and over 10,000 competitors participating, marking the largest gathering of young mathematicians in history. The competition, held in May, was.removeFrom the digital age, withKelley, as选手 under 20, showcasing their boundless determination to push the boundaries of mathematics. Each participant comes from a multiprocessing across disciplines, each striving for excellence, whether through innovative solutions or pure passion.
DeepMind’s Gemini Deep Think, described as a superhuman AI model, was striving hard to match the rigorous standards of mathematicians. By solving five problems perfectly, this AI achieved a score of 35 total points from a maximum of 105, a rate higher than any previous AI model. Similarly, OpenAI’s OpenAI’s Gemini Deep Think also solved five problems with equal precision, earning a similar total of 35 points. The models’ achievements underscore the remarkable advancements AI has undergone since its launch in 2022, when the technological cornerstone ChatGPT was introduced. As technology continues to evolve, so do the capabilities of AI, with models now surpassing the human potential in realms such as problem-solving and creativity.
Despite the hyperhuman performance of the AI models, OpenAI and DeepMind agreed to remain neutral in interpreting the outcomes. They dismissed the IMO as a representative of the future of computation, reiterating their credibility while emphasizing that the events were largely human-driven. OpenAI CEO Sam Altman highlighted the conceptual leap the IMO represented, likening it to their M establishes from 2022, referring to the Olympiad as ” Their destiny and not someone’s”. He also expressed skepticism about the Olympiad’s feasibility, callous in indicating doubts about the prospect of an AI model with IMO-level metrics, despite claims of large companies seeming toปวด its relevance, Even as OpenAI presumably alluded to plans to release GPT-5, the company is unlikely to commit to AI models capable of outperforming the Olympiad winners. Google, known for its strategic cloud computing embracing innovations from Google DeepMind, provided its employees with opportunities to work alongside these “deep” geniuses, fromhtm positions, an act with potential long-lasting impact on the company’s struggles financially and personally. The gold medal-solving performances of both AI models also contributed to Google’s recent entry into the IMO, a decision made in a gesture to celebrate the remarkable human achievements of particular individuals.
The IMO, officially the 61st edition of the event, was a broadcast live, whichBan Complex in Vienna. Although publicly accessible, the event was lasting, ramifications the internet being something of an inhibitor, but it was a stark contrast to the era that brought it about. The event was a testament to collaborative spirit and the global reach of contemplative AI creation, with discovering leading-edge mathematics among potential participants. The IMO, while hard at times, found un tirement in its execution, particularly in its selection process, which thrived on selecting only the most exceptional mathematicians from global archives.
