Charles O’Neill of Baseten, a cloud computing company, using the new artificial intelligence model from Z.ai (source: video by Justin T. Gellerson).
China’s artificial intelligence industry is rapidly challenging the dominance of U.S. developers by releasing high-performance models that deliver competitive capabilities at a fraction of the cost. The latest example is Z.ai’s GLM-5.2, an open source model that approaches the performance of Anthropic’s newest systems while costing only about one eighth as much for certain tasks. Its release coincided with temporary restrictions placed on Anthropic’s most advanced models, prompting businesses to reconsider their dependence on a handful of premium AI providers, tells The New York Times.
The rapid adoption of GLM-5.2 highlights a broader shift in the global AI landscape. Chinese companies now account for six of the world’s top 10 AI models on major industry leaderboards, demonstrating that the performance gap between American and Chinese developers has narrowed significantly. The new model has gained particular popularity for software development and AI agents, intelligent assistants capable of carrying out tasks across multiple applications. Major cloud providers, including Microsoft and Amazon, already offer access to AI models developed by Z.ai, DeepSeek, MiniMax, and other Chinese startups.
Despite their technical strengths, Chinese AI models continue to face significant obstacles in the United States. Concerns remain over potential ties between developers and the Chinese government, data security, censorship, and export restrictions. Z.ai was added to the U.S. Commerce Department’s trade blacklist in 2025, and several of its investors have links to Chinese state organizations. Nevertheless, organizations can deploy the open source models on their own infrastructure, limiting direct data exposure while benefiting from lower operating costs.
The article also examines the growing dispute over AI development practices. Anthropic and OpenAI accuse some Chinese companies of using unauthorized data harvesting and model distillation to accelerate development, allegations that Chinese firms have not publicly addressed. At the same time, researchers note that distillation alone cannot produce frontier AI systems without significant engineering expertise and large-scale computing resources.
As governments debate regulation and open source policies, many experts warn that limiting open AI development in the United States could strengthen China’s position. Although export controls continue to restrict China’s access to advanced AI chips, analysts estimate that Chinese developers trail U.S. companies by only a few months. For businesses seeking lower costs and greater flexibility, the emergence of powerful Chinese AI models has introduced meaningful competition into a market that until recently was dominated by a small group of American companies.