Chinese AI models emerge as competitive, low-cost alternatives to US peers

Chinese artificial intelligence models are emerging as highly capable, low-cost alternatives to US-based frontier models. Industry leaders, including Nvidia CEO Jensen Huang, have noted the growing capabilities of Chinese open-source and open-weight models, which include Alibaba’s Qwen, Moonshot AI’s Kimi, DeepSeek, and Z.ai’s GLM.

Alibaba Cloud recently updated its Qwen3.8 Max model to enhance coding, visual analysis, and tool-use capabilities. The updated model features a context window of one million tokens and offers both thinking and non-thinking modes to balance reasoning depth with economic efficiency. Alibaba has also reduced pricing for the model in the Singapore market by 20% compared to its predecessor.

While these models offer advantages in cost and performance, experts urge caution regarding security and data privacy. The Cloud Security Alliance has highlighted the importance of open-weight models as essential fallbacks for defenders during security incidents, noting that the same guardrails protecting frontier models can sometimes hinder defensive investigations.