Will China close the technology gap with the US?
China's AI models are rapidly closing the gap with US rivals, challenging the long-held US lead in artificial intelligence. Startups like Moonshot AI, DeepSeek, and Z.ai have released models such as Kimi K3, R1, GLM-5.2, and GLM-5.3, demonstrating capabilities comparable to Anthropic's Claude and OpenAI's GPT. Moonshot AI's Kimi K3 briefly ranked third globally in model intelligence in July. This development occurs despite US-led restrictions on advanced chip access for China.
The narrative of China's AI lag is outdated. Beijing-based Moonshot AI's Kimi K3 model briefly hit third globally in model intelligence in July. DeepSeek's V4 model, previewed in April, shows significant improvements in reasoning and agentic capabilities. Alibaba's Qwen3.8-Max also challenges top US models. The US lead, while still present in some areas, is fragile and narrowing.
This shift means Asia's AI development will not be a simple story of US dominance. Chinese models offer cost-effective alternatives, as seen with DeepSeek's R1. This competition could accelerate AI innovation across Asia, pushing regional players to differentiate. It also forces a re-evaluation of the impact of US chip restrictions on China's long-term AI trajectory.
The critical test for China's AI will be sustained performance and global adoption. Kimi K3's drop from third to ninth place shows the fluidity of the rankings. Watch whether Chinese models can maintain top-tier capabilities against rapid US advancements. The upcoming Xi-Trump meeting in Washington this month will also shape the regulatory environment for AI development.
