The United States and China are locked in a race to build ever-more powerful artificial intelligence systems, and experts say China has been rapidly closing the gap this year even as U.S.-led restrictions have blocked its access to some advanced chips and manufacturing equipment.
Experts and recent reports, including one by the Stanford Institute for Human-Centered Artificial Intelligence, say China has emerged as an AI counterweight to the U.S. and appears to be erasing the U.S. lead. “The gap between U.S. and Chinese models is narrow and fragile,” said Samm Sacks, a senior fellow at the Johns Hopkins School of Advanced International Studies’ Institute for America, China, and the Future of Global Affairs.
Chinese AI models pushing capabilities
Beijing-based Moonshot AI drew attention with Kimi K3, which shortly after release was ranked No. 3 globally in model intelligence in July by benchmarking platform Artificial Analysis, behind Anthropic’s Claude Fable 5 and OpenAI’s GPT-5.6 Sol. Kimi K3 now ranks ninth, trailing newer frontier models from Anthropic, OpenAI, Meta and SpaceXAI.
Startup DeepSeek demonstrated cost-competitive performance with its R1 model, released in January 2025, and in April previewed a more powerful V4 model it said improves knowledge, reasoning and so-called “agentic” capabilities — the ability to plan and execute complex steps autonomously under limited human guidance.
Z.ai rolled out GLM-5.2 in June and followed with GLM-5.3 in August, the company said, with improvements in coding and agentic functions that prompted comparisons to advanced Anthropic and OpenAI models. In July, Alibaba previewed Qwen3.8-Max, which the company described as its most powerful model in the series and a contender to models such as Claude Fable and GPT.
Security concerns in Beijing and accusations from Washington
Chinese officials have voiced their own concerns about AI risks. In an article published Sunday, Chen Yixin, head of China’s state security ministry, warned AI could threaten political, institutional and ideological security if used by “hostile forces and individuals with ulterior motives,” and suggested AI-generated deepfakes could spread “political rumors” at scale. Chen also cited capabilities of Anthropic’s Claude and OpenAI’s GPT as examples of tools that could be used to identify and “weaponize” cybersecurity vulnerabilities and enable complex hacking tasks.
At the same time, U.S. leaders have emphasized maintaining an AI edge. “Whoever wins AI, wins,” President Trump told reporters on Sunday. U.S. agencies have accused Chinese developers of engaging in “industrial-scale distillation,” a practice in which models are trained to copy or imitate more powerful systems. Anthropic has accused Chinese startups DeepSeek and Moonshot of routing a number of their user requests to its Claude models.
In a Sept. 8 report, the FBI, the National Security Agency and the Cybersecurity and Infrastructure Security Agency said Chinese AI developers extracted capabilities from some of the most advanced U.S. models in what the agencies described as “aggressive, malicious, and targeted distillation activities.” Beijing rejected those claims, saying distillation is commonly used across the industry, asking the United States to refrain from “unfounded accusations” and accusing the U.S. of pursuing a “monopoly of the AI industry.”
Lizzi C. Lee, a fellow at the Asia Society Policy Institute’s Center for China Analysis, noted the allegations remain contested but carry strategic weight. “These are still contested allegations at this point but I think the strategic implication is already important,” she said. Lee added that while the U.S. has focused on controlling inputs into frontier AI — especially chips — it now faces questions about whether the outputs of frontier AI effectively transfer capability. “The U.S. labs spend enormous amounts of compute pushing out the frontier. But Chinese developers can learn from those systems,” she said. “And Chinese companies then [can] release very capable open-weight models that spread globally.”
Top Chinese leaders have called for accelerating AI development and pursuing technological self-reliance while also stressing the need to control risks and seek a global framework for AI governance. The future direction of AI, experts say, will depend significantly on the actions and responses of both China and the United States.
