On September 21, 2026, amidst the scientific research achievements exhibition at the High-Quality Development of Teacher Education and the 70th Anniversary Celebration of Qinghai Normal University, China proudly unveiled its groundbreaking creation—the Zhida AI Input Method. This innovative Tibetan multilingual full-modal AI input method, accompanied by a matching set of Tibetan fonts, was officially introduced to the public. Developed by the National Key Laboratory of Tibetan Intelligence and powered by the cutting-edge Zhida Large Model technology, this input method boasts comprehensive coverage across all terminals and supports multimodal input.

Adhering to the national standard Tibetan keyboard layout, it seamlessly integrates Tibetan Latin transcription capabilities. Its voice input feature is particularly noteworthy, as it encompasses the three major dialects of Tibetan and can accurately process mixed Chinese-Tibetan-English speech. Additionally, the OCR recognition function can decipher mixed Chinese-Tibetan-English text with precision. The input method's vocabulary is extensive, encompassing terms from 23 academic disciplines, and it can intelligently synchronize with users' input habits for a more personalized experience. The accompanying Tibetan font set is available in four distinct styles, catering to diverse aesthetic preferences.

Furthermore, the conference also showcased the remarkable Zhida Large Model. This model comes in nine versions, with parameter sizes ranging from 800 million to an impressive 100 billion. Among them, three main models have been deployed online, offering mutual translation capabilities among 56 languages, as well as image and video understanding functionalities. This versatile model serves multiple fields, demonstrating its broad applicability and potential.