Alibaba Unveils Qwen3.8 AI Model With 2.4 Trillion Parameters

Alibaba Claims Runner-Up Position in AI Model Race

Alibaba Group Holding has introduced a preview of Qwen3.8-Max, its most powerful artificial intelligence model to date, featuring 2.4 trillion parameters. The company positioned the system as competitive with leading Western AI models, stating it ranks "second only" to Anthropic's Claude Fable 5.

The preview version is now accessible through Alibaba's Token Plan subscription service and its Qoder and QoderWork agentic platforms, according to an announcement the company made on social media over the weekend.

China's Multi-Trillion Parameter Milestone

The release marks a significant milestone in China's domestic AI development efforts. Parameters serve as a key metric for measuring AI system complexity, with larger numbers typically correlating to more capable models. Alibaba's 2.4 trillion parameter count places it among the most sophisticated AI systems currently available.

The announcement came during the World AI Conference in Shanghai, China's premier artificial intelligence industry event. The timing underscores the competitive intensity among Chinese AI laboratories as they work to match capabilities demonstrated by U.S.-based companies.

Alibaba's announcement followed closely on the heels of another major Chinese AI development. Just days earlier, Beijing-based Moonshot AI revealed Kimi K3, which features 2.8 trillion parameters and currently holds the distinction of being the world's largest open-source AI model.

Why it matters

The rapid advancement of Chinese AI models into the multi-trillion parameter range demonstrates that domestic labs are closing the technical gap with Western competitors faster than many observers anticipated. For enterprise technology leaders, this shift signals an increasingly multipolar AI landscape where competitive alternatives to U.S. models are emerging with comparable capabilities. Organizations operating in or with China may soon have access to locally developed models that rival the performance of systems from OpenAI, Anthropic, and Google, potentially reshaping procurement decisions and deployment strategies in the world's second-largest economy.

Competitive Context

By benchmarking itself against Anthropic's Claude Fable 5, Alibaba is explicitly positioning Qwen3.8 within the tier of frontier AI models. Anthropic's Claude family has gained recognition for strong performance across reasoning, coding, and natural language tasks, making it a meaningful comparison point for assessing model capabilities.

The preview designation suggests Alibaba may refine the model further before a full commercial release, a common practice that allows developers to gather real-world feedback while managing expectations around performance and reliability.

These details were first reported by the South China Morning Post.