According to the latest AI Index Report released by Stanford University's Human-Centered Artificial Intelligence Institute, China is rapidly catching up with the US in developing cutting-edge artificial intelligence (AI) models.
The report found that in 2024, the US maintained a lead with 40 major AI models, followed by China with 15 and France with three. Despite the gap in quantity, Chinese models are closing in on performance.
The performance difference between US and Chinese models on key benchmarks has narrowed significantly. On the Massive Multitask Language Understanding (MMLU) test, the average accuracy gap shrank from 17.5% in 2023 to just 0.3% in 2024. The rise of Chinese developers such as Alibaba, ByteDance, Tencent, DeepSeek, and Zhipu AI—now among the global top 15 AI model creators—has played a key role in this progress.
DeepSeek's V3 model, launched in December 2024, was specifically highlighted for its ability to deliver high performance while using significantly fewer computational resources than many of its international peers.
China also continues to lead in AI research publications and patent filings. In 2023, it accounted for 69.7% of all global AI patents, compared to the US's 14.2%, a steep drop from 42.8% in 2015. The East Asia and Pacific region collectively contributed 82.4% of the world's AI patents last year.
While Chinese researchers produced 23.2% of AI research papers and garnered 22.6% of total citations, US-based research retained greater global influence, especially in studies cited over 100 times.
Despite China's rapid technical progress, the Stanford report warned of potential roadblocks, particularly in funding. US government restrictions on outbound investment in Chinese AI startups have impacted capital flows. In 2024, private AI investment in the US soared to US$109.1 billion, nearly 12 times higher than China's US$9.3 billion.
The report underscores China's growing AI capabilities but suggests that regulatory headwinds and funding disparities may slow its momentum in the years ahead.
Article translated by Jingyue Hsiao and edited by Jerry Chen
