Chinese 'Alibaba' launches AI model that combines text, images, video, and task execution

Chinese company Alibaba has launched its new AI model 'Qwen 4 Max', which features 2.4 trillion parameters and integrates text processing, image and video understanding, and complex task execution using external tools, with improved performance and reduced resource consumption.

Chinese company Alibaba has launched its new AI model 'Qwen 4 Max', which features 2.4 trillion parameters and integrates text processing, image and video understanding, and complex task execution using external tools. The launch follows a limited beta release earlier in November, aiming to expand AI capabilities from generating answers and content to executing multi-step tasks including search, information analysis, and file and project preparation. According to Qwen Cloud documents, the model supports handling text, images, and video clips, along with web search, document reading, digital tool invocation, and structured output generation. The company stated that the new model has enhanced capabilities for professional and scientific tasks, as well as improved task execution efficiency compared to the previous version. Dow Jones reported that Alibaba said the model contains 2.4 trillion parameters and uses fewer computing resources than its predecessor to accomplish the same tasks, thereby improving operational efficiency. The company added that the model achieved top scores in benchmarks measuring programming skills, tool-based task execution, and visual reasoning.

Chinese 'Alibaba' launches AI model that combines text, images, video, and task execution