Google launches WeatherNext 3 and wires it into Search, Maps and Gemini

Google DeepMind and Google Research released WeatherNext 3 on Wednesday, describing it as their most advanced and accurate global weather AI model. The company says the model now incorporates real-time satellite data, refreshes hourly, runs at higher resolution, delivers more precise precipitation forecasts and adds clean energy variables. It is integrated across Google Search, Gemini, Maps, the Google Maps Platform and Google Cloud, where users and researchers can access the data for their own projects. Google senior staff engineer Samier Merchant told TechCrunch this is the first time some of the core variables will feed and power many Google products. According to TechCrunch, WeatherNext 3 has proven the most accurate among leading contenders on Operational WeatherBench, a comparison utility built by the startup Brightband that examines metrics such as temperature, wind speed and humidity. On that benchmark it beats other deep-learning models from Google, Microsoft, Nvidia and the European Centre for Medium-Range Weather Forecasts, as well as traditional forecasts from the U.S. National Weather Service and ECMWF. Google says the improvements matter for agriculture, renewable energy and daily planning. The release continues a shift in meteorology toward learned models that are faster and cheaper than the physics simulations run on government supercomputers.