Stanford AI Index: US-China Model Gap Narrows to 2.7 Points

Stanford University's Institute for Human-Centered Artificial Intelligence released its ninth annual AI Index Report on Monday, revealing that the performance gap between leading U.S. and Chinese AI models has nearly closed, with the top U.S. model surpassing its closest Chinese rival by just 2.7 percentage points as of March 2026.

The report, drawing on data from sources including the community-driven Arena platform, highlights a competitive landscape where U.S. and Chinese models have frequently swapped positions atop global benchmarks since early 2025. For instance, China's DeepSeek R1 matched top U.S. models in February 2025, and subsequent releases from companies like Alibaba, Zhipu and MiniMax have sustained high rankings. While the U.S. produced 50 notable AI models in 2025 compared to China's 30, and led in private investment at $285.9 billion versus China's $12.4 billion, the latter's government-backed funding since 2000 totals an estimated $184 billion. The U.S. also dominates in high-impact patents, infrastructure with 5,427 data centers, and new AI startups, numbering 1,953 last year.

China, however, excels in scale, accounting for 23.2 percent of global AI...