The AI race enters a new phase with Alibaba’s latest Qwen model
Alibaba Group on Monday officially unveiled its most powerful AI model to date, the Qwen 3.8 Max. After the overnight launch, Alibaba’s New York-listed shares climbed 4.5% on Monday, while its shares rose 7% on the Hong Kong exchange.
The model contains 2.4 trillion parameters and is competitive with Anthropic’s AI models & leading AI developers. Also supports a context window of up to 1 million tokens, which means it can understand and process extensive documents and media.
The announcement comes just weeks after domestic rival Moonshot AI introduced Kimi K3, a model featuring 2.8 trillion parameters that is widely regarded as one of China’s largest AI models.
Alibaba shared results comparing the performance after testing and showing the Qwen AI model scores better than Anthropic’s Fable 5. The model ranks second in the Vision Arena and fifth in the Text Arena, highlighting its strong capabilities across both multimodal and text-based AI tasks.
As per the Qwen team, it delivers “comprehensive improvements across coding, work, research, and long-horizon tasks” with minimal supervision. They allow the system to recognise patterns, understand language, generate responses and solve problems.
You cannot directly download the Qwen 3.8-Max model weights. However, Alibaba has confirmed that the model will be released as open weights next week. Once available, developers will be able to download the weights from platforms such as Hugging Face and the official Qwen AI website.
Developers can access the model via Alibaba Cloud using their Token Plan subscription.
