Singapore's National AI Initiative Abandons Meta Model

It was learned on November 25th that Singapore's National AI Program (AISG) is undergoing a major strategic adjustment. In its latest large language model project for Southeast Asian languages, it has abandoned the Meta model and shifted to Alibaba's Tongyi Qianwen Qwen open-source architecture. This marks a key expansion of Chinese open-source AI models in the global influence landscape.

It is reported that the "Qwen-SEA-LION-v4" model, announced by AISG on November 25th, quickly climbed to the top of an open-source benchmark measuring Southeast Asian language capabilities. This shift aims to address a long-standing pain point: previous open-source models, represented by Meta's Llama series, performed poorly when processing regional languages such as Indonesian, Thai, and Malay, severely constraining the development efficiency and performance of localized AI applications.

This change also highlights the rise of Alibaba's Tongyi Qianwen in the global open-source AI field. Since its full open-sourcing in 2023, Alibaba's Qwen has surpassed models like Llama and DeepSeek to become the world's strongest and most widely used open-source large language model. As of now, the global download count for the Qwen model series has exceeded 600 million.

While accumulating a high reputation within the global technical community, Alibaba is also expanding into the AI consumer market. Its Qianwen App, launched based on the Qwen model, surpassed 10 million downloads in its first week, becoming the fastest-growing AI application in history.