Alibaba shares rose on Monday after the company introduced its biggest artificial intelligence model so far. This move adds to evidence that China’s focus on open AI development is putting more pressure on American model providers.
American depositary receipts went up 4.6% after Alibaba launched its newest model, Qwen 3.8-Max, which has 2.4 trillion parameters. This number shows the model’s size and complexity. Alibaba said the model performed about as well as the latest models from OpenAI and Anthropic on coding tests.
Alibaba also announced it will open-source the model’s weights at this scale for the first time. This means developers can download the parameter values and use or change the model on their own hardware. In contrast, U.S. companies usually keep their top models private and charge higher subscription fees for access.
William Blair pointed out that Chinese open-weight models usually cost about one-tenth as much as American ones. This lowers the premium that developers of proprietary models can charge, especially when a slightly less advanced model is good enough. Alibaba did not say what hardware it used to train the model or how much it cost to develop.
This announcement comes after a similar move last month by Moonshot AI, a Chinese lab supported by Alibaba. Its Kimi-K3 model surprised the market by doing better than top models from Anthropic and OpenAI in various coding tasks, according to independent evaluator Arena AI.
