Tibet moves from AI pilots to full-scale Tibetan-language large model development

Summary

China News Service reports that Tibet has achieved a major milestone in developing a trillion-parameter Tibetan-language base model "Sunshine Qingyan" V1.0, marking a shift from limited AI applications to systematic large-model R&D in the region. Local startup teams have also built the DeepZang Tibetan large model—now integrated into the DeepSeek ecosystem—with plans to release a public app by year-end, which could significantly expand AI access for Tibetan speakers in government, community services and education.