For the first time, artificial intelligence systems have achieved a perfect score at the International Mathematical Olympiad (IMO), the world’s most prestigious competition for young mathematicians. Two Chinese companies, Huawei and Xiaohongshu, announced that their models each solved all six problems correctly at this month’s event in Shanghai, matching the performance of only seven of the 666 human contestants. The milestone, confirmed under the IMO’s official judging process, marks a step beyond the gold-medal-level scores attained by Google and OpenAI models in 2025, which had fallen short of the few perfect human results.

The speed of progress is striking. At the 2024 IMO, Google’s system reached a silver-medal equivalent, solving four problems over two to three days. This year, the Chinese models were given the problems only after the human exam ended and had to submit answers within a set time limit, with no human intervention permitted. A partner at the US investment firm Menlo Ventures, Deedy Das, reported on LinkedIn that he independently tested four cutting-edge models from OpenAI, Anthropic, the startup Axiom Math, and China’s Moonshot AI on the same questions; all scored 42 out of 42 points. Das wrote that “the frontier of AI has officially moved well past IMO math.”

The achievement lands amid a broader push by Chinese technology firms to convert AI prowess into global commercial reach. Customs data for the first half of 2026 show integrated-circuit exports nearly doubling year on year to US$177.3 billion, while exports of automatic data-processing equipment rose over 41 per cent to US$138.1 billion. Beyond hardware, Shanghai-based AI lab MiniMax disclosed that 73 per cent of its 2025 revenue—US$57.7 million—came from outside mainland China, up from US$21.3 million a year earlier, serving more than 214,000 enterprise customers and developers across over 100 countries. Its flagship model became the first Chinese system on the OpenRouter platform to exceed 50 billion in daily token consumption.

Other AI labs around the world are expected to publish their own IMO results in the coming days, and the contest’s organisers have not yet commented on whether the perfect scores will be formally recognised as equivalent to human medals. The development intensifies the race to demonstrate advanced reasoning capabilities, while the export figures and MiniMax’s revenue trajectory suggest that Chinese firms are simultaneously building international software and cloud-based AI businesses alongside their hardware supply chains.