Google unveils WeatherNext 3 for faster, higher-resolution forecasts
Google DeepMind and Google Research have introduced WeatherNext 3, an AI model designed to forecast atmospheric conditions with greater detail and frequency. Google says the system will contribute to weather information shown in Search, Google Maps and Gemini, and will also be offered to users and researchers through its cloud platforms.
In tests using Brightband’s Operational WeatherBench, WeatherNext 3 was the most accurate of the leading systems assessed. The comparison covered temperature, windspeed and humidity, and included deep-learning models from Google, Microsoft, Nvidia and the European Center for Medium-Range Weather Forecasting, as well as traditional forecasts from the US National Weather Service and ECMWF.
The model can forecast key variables at a 5km resolution, produce hourly predictions and showed a 60% improvement in rain evaluations over WeatherNext 2. It has 2.4 times more parameters than its predecessor and can use hourly satellite observations collected in real time. Google says it is the first AI model to directly incorporate raw observations for a high-resolution global forecast, although WindBorne says its WeatherMesh 6 has used raw observations since late 2025. Both systems still rely on national weather datasets.
Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.