U.S. export bans on AI accelerate China’s push to close the gap

A new generation of Chinese artificial intelligence models is eroding the lead long held by American labs, arriving at a moment when U.S. policy decisions may be inadvertently accelerating the shift.

GLM-5.2 Emerges as a Credible Challenger

Z.ai, the Chinese startup also known as Zhipu AI, has released GLM-5.2, an open-weight model that early benchmarks place within a single percentage point of Anthropic's Claude Opus 4.8 on agentic coding evaluations — at roughly one-sixth the API cost of leading U.S. models. The model features a one-million-token context window and is released under the permissive MIT license, allowing anyone to download and deploy it on private infrastructure.firstpost+3

The release has already drawn interest from major Western platforms. Zoho founder Sridhar Vembu noted on X that companies including Databricks and Microsoft ↘ 0.45% are embracing Chinese open-source AI not for geopolitical reasons but because it offers lower costs and greater deployment control. For most practical tasks below the absolute frontier, Chinese models are now "legitimate first choices," according to analysis drawing on the Stanford AI Index.youtube+1

Cybersecurity Parity With Anthropic's Mythos

On June 24, Zhou Hongyi, founder of 360 Security Technology, unveiled two AI security tools at the ISC.AI 2026 conference in Beijing. Tulongfeng is designed to automatically discover software vulnerabilities, while Yitianzhen automates cyber defense and incident response. Zhou described the tools as China's answer to Anthropic's Mythos 5, the cybersecurity-focused model that the Trump administration recently banned from export.forklog+1

Tulongfeng has already identified 3,432 software vulnerabilities, with 105 confirmed by Chinese regulators and several classified as highly critical. The Wall Street Journal News Corp ↗ 0.39% reported that Chinese AI systems have matched Mythos's performance in some cybersecurity scenarios, though they still trail Anthropic and OpenAI in other domains.wsj+1

U.S. Policy Vacuum Raises Stakes

The advances arrive as the Trump administration's restrictions on Anthropic's most powerful models have created what analysts describe as an opening for Chinese competitors. TechCrunch reported that the export ban on Mythos and Fable 5, imposed roughly two weeks ago, has spurred Asian startups to rush out alternatives. CNBC characterized the administration's AI crackdown as "a gift to China".cnbc+1

With Chinese open-weight models now freely available worldwide while American frontier systems face export controls, developers in India and elsewhere are turning to what the Economic Times called "fine China" — capable models without access restrictions. The gap between the top U.S. model and its Chinese equivalent stood at just 2.7 percent on aggregate benchmarks as of March 2026, according to the Stanford AI Index.economictimes+1