Data from Artificial
Analysis, a leading global independent AI model evaluation and comparison platform, shows that as of June 16, 2026, the intelligence gap between cutting-edge AI models from China and the US is rapidly narrowing. The chart compares the highest-performing models from each country each month, with the blue line representing the US model, the red line representing the Chinese model, and the gray area representing the percentage of the Chinese model relative to the US model.
The chart shows that from the end of 2023 to 2024, the US model was significantly ahead, with the Chinese model only reaching approximately 50%–60% of the US's leading level. However, after 2025, the Chinese model index accelerated its upward trend, reaching nearly 90% of the US's leading level by mid-2026.
This means that AI competition is shifting from "absolute US leadership" to "competition between China and the US in the same tier." For the market, the key change is that even facing limitations in high-end computing power, Chinese models are rapidly catching up through engineering optimization, model efficiency, and application iteration, weakening the framework of simply measuring AI competitiveness by computing power scale.
If this trend continues, the technological premium of US AI companies may face reassessment, while the narratives of domestic computing power, cloud services, edge AI, and industry applications will be further strengthened. Future AI investment may no longer focus solely on model parameters and training scale, but rather on who can achieve commercialization at a lower cost and faster speed.
