China launches first tibetan multilingual AI input method with voice and image support

On September 20, 2026, China unveiled its first Tibetan multilingual full-modal AI input method, Zhida AI Input Method, during an event marking the 70th anniversary of Qinghai Normal University. Developed by the National Key Laboratory of Tibetan Intelligent Technology, the input method supports text, voice, and image input across devices like phones and computers. It includes a Tibetan font set with four styles and is compatible with multiple dialects of Tibetan. Professor Duola from Qinghai Normal University highlighted that the input method is based on the Zhida large model, adhering to national Tibetan keyboard standards and supporting Latin transcription. Voice input covers three major Tibetan dialects and mixed Chinese, Tibetan, and English. OCR features allow recognition of mixed texts, aiding in document digitization and office tasks. The input method's dictionary includes terminology from 23 fields and adapts to user input habits. The Zhida large model, with nine versions ranging from 800 million to 100 billion parameters, supports 56 language translations and image/video understanding, applicable in education, research, and government sectors.