Stanford University's latest AI Index Report reveals a dramatic reshaping of the global artificial intelligence landscape, with China rapidly closing the performance gap with the United States in developing advanced AI systems.
The eighth edition of this authoritative report, produced by Stanford's Institute for Human-Centered Artificial Intelligence (HAI), shows that while US-based institutions still lead in quantity—producing 40 notable AI models in 2024 compared to China's 15—the quality difference has narrowed substantially. On major benchmarks like MMLU (Massive Multitask Language Understanding) and HumanEval, the performance gap between top US and Chinese models has shrunk from double-digit percentages in 2023 to near parity in 2024.
Perhaps most striking is the head-to-head comparison: in January 2024, the top US model outperformed the best Chinese model by 9.26%, but by February 2025, this advantage had dwindled to just 1.70%. Meanwhile, China continues to dominate in AI publications and patents, accounting for nearly 70% of all AI patent grants globally.
The report also highlights a concerning trend in AI evaluation: many benchmarks have become 'saturated' as AI systems achieve such high scores that the metrics no longer effectively differentiate between models. This saturation spans multiple domains including general knowledge, reasoning about images, mathematics, and coding, forcing researchers to develop more challenging evaluation frameworks.
Beyond the US-China competition, the report documents broader global trends: AI is becoming more efficient and affordable, with inference costs for high-performance models dropping 280-fold in just 18 months. However, AI-related incidents rose by 56.4% in 2024, underscoring the growing need for responsible AI guardrails as these technologies become more widespread.
The findings suggest that the international AI landscape is entering a new phase of intensified competition, with significant implications for technological leadership, economic advantage, and national security in the coming years.