Skip to content

China catches up in the race of developing advanced AI models

Rapid advancements in AI by China increasingly narrow the performance gap with the U.S. in language tests, reducing differences from 17.5% to a mere 0.3%, sparking worries about America's long-standing dominance.

China Progresses in AI Model Advancements, Narrows Gap with Leaders
China Progresses in AI Model Advancements, Narrows Gap with Leaders

China catches up in the race of developing advanced AI models

In the realm of artificial intelligence (AI), the global race for dominance continues, with the United States and China leading the pack. However, as of 2024 and early 2025, China has made significant progress in developing artificial intelligence models but remains behind the United States in overall AI capabilities, though the gap is closing rapidly.

Key details include:

  • Model Development and Quantity: In 2024, the U.S. had about 40 notable AI models, while China had 15, indicating a sizable but decreasing gap in quantity and quality of AI models between the two countries.
  • Technological Progress: China’s AI models are rapidly improving in key capabilities such as general reasoning, coding, and scientific problem-solving, showing significant year-over-year gains. For instance, coding accuracy jumped from 4.4% in 2023 to 71.7% in 2024, narrowing the quality gap with U.S. models.
  • Policy and Strategic Focus: China pursues a dual strategy balancing growth with increased control, investing heavily in AI infrastructure, regulations, and self-reliance, especially amid restricted access to advanced foreign AI hardware due to export controls.
  • Domestic Industry Growth: China’s AI sector exceeded 700 billion yuan (~$97.5 billion) in scale in 2024, maintaining 20%+ annual growth for several years. Key domestic players like Alibaba, Tencent, and Huawei drive innovation across the AI stack—from chips to large language models.
  • Global Ambitions: China announced a comprehensive global AI governance action plan in mid-2025 and proposes an international AI cooperation organization to promote coordination and showcase its AI advancements, reflecting its ambition to influence global AI development and standards.
  • Overall Assessment: Experts note that although the U.S. remains the global leader, China’s rapid advancements and strategic focus mean it could catch up or compete at near parity if the U.S. slows down. The competition spans not just technology but also AI governance and international cooperation.

Meanwhile, concerns over the proximity of Taiwan to China have raised fears that Beijing could stack its tech deck against Washington. The U.S. has imposed export controls on high-tech chips, limiting access for Chinese firms like Huawei. Additionally, some states have banned the Chinese AI app DeepSeek on government devices due to concerns over its use for surveillance and censorship.

The June Fox News Poll showed that most registered voters had little confidence in the government's ability to properly regulate AI, reflecting the complexity and urgency of the issue. NVIDIA Founder and CEO Jensen Huang stated that controlling the standards is crucial to ensure every AI developer and country is built on the American tech stack.

In the event of a military attack on Taiwan, the sensitive equipment could be de-calibrated or destroyed, further complicating the technological landscape. The Taiwan Semiconductor Manufacturing Company leads the field in the final layer of the technology stack, several years ahead of the Chinese equivalent, SMIC.

The Trump Administration has allowed NVIDIA chips to enter the Chinese market again, competing against Beijing's fast-advancing semiconductors. In a more advanced language test, Chinese Platform DeepSeek-R1 posted a higher score than any American-made model. In General Reasoning testing, the U.S.'s 13.5% advantage fell to 8.1% in the same period. In 2024, the U.S.'s advantage over China in language understanding tests narrowed from 17.5% in 2023 to 0.3%.

David Sacks, the White House AI and Crypto Czar, stated that China is showing the most improvement in the AI model layer. Models are frequently assessed in various tests, including languages, math problems, general reasoning, and coding. DeepSeek, a Chinese AI model, is estimated to have more than 20 million daily users.

As the race for AI dominance continues, both the U.S. and China are pushing the boundaries of technological innovation, with China rapidly closing the gap. The competition is not just about technology but also governance and international cooperation, making it a complex and dynamic landscape.

  1. The steady progress in artificial intelligence (AI) model development by China, such as DeepSeek, which has an estimated 20 million daily users, is causing concern for the US, as the gap between the two nations is closing remarkably in both quantity and quality.
  2. In the realm of AI, not only is technology innovation at stake, but also the effort to establish global standards is critical, as expressed by NVIDIA Founder and CEO Jensen Huang, who emphasizes the importance of controlling the standards to ensure every AI developer, including China, is built on the American tech stack.

Read also:

    Latest