Alibaba unveils Qwen3.8-Max AI model with 2.4 trillion parameters in bid to challenge US dominance

1 hour ago 22

Alibaba just dropped its most powerful AI model ever, and the timing feels like a statement. Qwen3.8-Max, announced on August 3, packs 2.4 trillion parameters into an open-weight multimodal system designed to go toe-to-toe with the best that Silicon Valley and Beijing have to offer.

Alibaba’s stock responded accordingly, climbing as much as 3.4% in a single trading session.

What makes Qwen3.8-Max different

At 2.4 trillion parameters, the model sits just below Moonshot AI’s Kimi K3, which boasts 2.8 trillion parameters. But it’s competitive with Anthropic’s Fable 5 and comparable offerings from OpenAI, putting Alibaba squarely in the conversation with the world’s leading AI labs.

Alibaba is making Qwen3.8-Max open-weight, meaning developers anywhere in the world can download the model’s parameters and build on top of them. OpenAI and Anthropic keep their models locked down behind APIs and subscription tiers.

The model is set to be available for widespread developer access within one week of the announcement.

In performance testing, Qwen3.8-Max showed particular strength in coding tasks and visual reasoning. One notable benchmark: the model completed a software engineering project in 16 days.

The bigger picture: China’s AI counteroffensive

Alibaba’s release is part of a calculated response by Chinese tech firms to US export restrictions on advanced chip technologies. Companies like Alibaba, DeepSeek, and Moonshot AI have doubled down on innovation, with the open-weight strategy allowing Chinese firms to attract global developers who might otherwise default to US platforms.

The Qwen series has been central to Alibaba’s AI ambitions for some time, progressing through multiple iterations focused on open-source accessibility and multilingual capabilities. Qwen3.8-Max represents the most ambitious step yet in that lineage.

Domestic competition is fierce too. DeepSeek made waves earlier with its own efficient AI models, and Moonshot AI’s Kimi K3 currently holds the parameter crown at 2.8 trillion.

What this means for investors

One risk to flag: open-weight doesn’t mean open-source in the traditional sense. The license terms governing how developers can use Qwen3.8-Max will matter enormously. Previous open-weight releases from various companies have come with restrictions that limit commercial use or geographic availability. Investors should watch the actual license terms rather than taking “open” at face value.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.

Read Entire Article