Alibaba unveils Qwen3.8-Max, a 2.4 trillion parameter AI model
Alibaba unveiled Qwen3.8-Max on Monday, a 2.4 trillion parameter AI model the company says is competitive with Anthropic’s latest system. The announcement sent Alibaba’s New York-listed shares up 4.5% in U.S. trading, with Hong Kong-listed shares rising 7%, as Chinese technology companies continue pushing to close the AI gap with U.S. rivals.
What Alibaba released
Qwen3.8-Max is the largest model in Alibaba’s Qwen family to date. It supports a context window of up to 1 million tokens, allowing it to process and reason over text equivalent to thousands of pages. Parameters are the numerical settings that shape how an AI model processes information and produces responses; a higher count generally indicates a larger and more capable model.
The model covers a broad capability range, including coding, general productivity, research, long-horizon tasks, and visual intelligence. Alibaba said Qwen3.8-Max is scheduled for release next week.
How it compares on benchmarks
Alibaba shared benchmark results showing Qwen3.8-Max matching or beating Anthropic’s Fable 5 on several tests. The model ranks second to Fable 5 in the Vision Arena and fifth in the Text Arena. The Anthropic comparisons position Qwen3.8-Max as one of the strongest non-U.S. models publicly measured against a leading Western system.
Autonomous coding and long-running tasks
Alibaba highlighted Qwen3.8-Max’s ability to operate autonomously for extended stretches. In an internal test, the model spent 16 days building and improving an AI coding tool on its own, writing code, running tests, fixing errors, and refining its output with little human input. The company framed the capability as useful for production-style coding projects that would normally require continuous developer oversight.
Beyond coding, Alibaba said the model can handle real-world workloads such as reviewing legal documents, conducting financial research, and producing architectural 3D modeling.
Visual intelligence and large-context understanding
The model’s visual capabilities, according to Alibaba, extend to parsing hundred-page documents, full television series, or 100-hour livestreams and turning the content into searchable, interactive knowledge hubs. The 1 million token context window underwrites these use cases by letting the model keep very large amounts of material in working memory.
How Qwen3.8-Max fits China’s AI race
Qwen3.8-Max arrives during a period of heavy investment by Chinese AI companies. Earlier in August, domestic rival Moonshot AI released Kimi K3, a model with 2.8 trillion parameters that is currently China’s largest publicly disclosed AI model. Alibaba’s 2.4 trillion parameter count places Qwen3.8-Max second in that domestic ranking but still among the largest models worldwide.
Chinese and U.S. labs have spent 2025 and 2026 trading positions on public leaderboards as both governments treat AI leadership as a strategic priority. Alibaba’s Qwen line has been one of the most active Chinese model families, with frequent updates aimed at enterprise and developer users.
Market reaction
Investors treated the announcement as a signal of competitive standing. The 4.5% gain in New York-listed shares and the 7% gain in Hong Kong on the same day marked one of Alibaba’s stronger single-session moves tied to AI news in 2026. The rally reflects both the scale of the model and Alibaba’s positioning as a credible alternative to U.S. frontier labs for buyers seeking non-U.S. suppliers.
FAQ
What is Qwen3.8-Max?
Qwen3.8-Max is Alibaba’s largest AI model to date, containing 2.4 trillion parameters and supporting a context window of up to 1 million tokens. It is scheduled for release next week.
How does Qwen3.8-Max compare to Anthropic’s models?
Alibaba shared benchmark results showing Qwen3.8-Max matching or beating Anthropic’s Fable 5 on several tests. The model ranks second to Fable 5 in the Vision Arena and fifth in the Text Arena.
What can Qwen3.8-Max do?
Alibaba says Qwen3.8-Max can code autonomously for extended periods, in one test spending 16 days building and improving an AI coding tool on its own. It can also review legal documents, conduct financial research, produce architectural 3D modeling, and parse hundred-page documents, full television series, or 100-hour livestreams into searchable, interactive knowledge hubs.
This article summarizes reporting from cnbc.com.