What happened
Alibaba Cloud partnered with Singapore's national AI programme AI Singapore to build the regional multilingual model Qwen-SEA-LION-v4 (released November 2025) based on Tongyi Qwen3-32B: post-trained on more than 100 billion tokens of Southeast Asian language data, covering Indonesian, Malay, Thai, Vietnamese, Burmese, Filipino, Tamil, and other languages; it offers 4-bit/8-bit quantised versions that can run on consumer-grade laptops, is freely available via Hugging Face, and has topped the SEA-HELM leaderboard among open-source models of the same class.
Localisation & innovation
A regional adaptation paradigm of joint post-training with the host country's national AI institution, led by local-language corpora; the shift of the SEA-LION v4 foundation model from a Western model to China's Qwen is seen as a landmark regional win for China's open-source ecosystem