China has unveiled an advanced artificial intelligence (AI)-powered large language model, Fenghe, specifically designed for meteorological applications. Launched on Tuesday by the China Meteorological Administration’s Public Meteorological Service Center, the model aims to provide smarter, personalized, and scenario-based weather services for both the public and various industries. According to Zhu Xiaoxiang, director of the center, Fenghe integrates general large language model capabilities with specialized meteorological knowledge, trained on 550 billion tokens of meteorological data to ensure accuracy. Users can access the model via a miniprogram or website, receiving immediate AI-generated responses to weather-related queries, ranging from disaster alerts to tailored suggestions for transportation, tourism, logistics, energy, and health. For example, in self-driving scenarios, Fenghe acts as a travel adviser, offering real-time recommendations such as avoiding flooded highways during storms or suggesting indoor destinations during severe weather. The system also connects to real-time meteorological data systems like Fenglei for nowcasting and Fengqing for short-range forecasting. Li Juanzi, a professor at Tsinghua University, emphasized Fenghe’s deep integration with real-world applications, aiming to meet public life, administrative, and emergency coordination needs. Wang Jianghai, deputy director of the Ministry of Emergency Management’s flood and drought disaster prevention department, highlighted the model’s role in enhancing disaster prevention and response capabilities, describing it as an expert adviser for flood control workers and a reliable weather companion for the public. The China Meteorological Administration plans to continuously evaluate and optimize Fenghe through a standardized testing system to ensure its long-term effectiveness.