Google Releases WeatherNext 3 AI Forecasting Model

Scientists at Google DeepMind and Google Research have released WeatherNext 3, a new artificial intelligence model for weather forecasting. Google says the model will feed weather information shown in Search, Google Maps, and Gemini, and will be available to users and researchers through Google’s cloud platforms.

WeatherNext 3 was rated the most accurate model among leading contenders tested on Operational WeatherBench, a comparison tool built by Brightband. The evaluation covers temperature, windspeed, and humidity. The model outperformed deep-learning models from Google, Microsoft, and Nvidia, as well as forecasts from the European Centre for Medium-Range Weather Forecasts and the US National Weather Service.

Researchers say WeatherNext 3 can forecast key variables at a 5km resolution. Its rain evaluations improved by 60% over WeatherNext 2, and it can produce hourly forecasts instead of the standard six-hour interval.

The model has 2.4 times more parameters than its predecessor. It can also target forecasts to specific weather stations and ingest real-time satellite data collected hourly. Google calls it the “first” AI model to directly incorporate raw observations for a high-resolution global forecast, although WindBorne says its WeatherMesh 6 model has used raw observations since late 2025. Both models still rely on national weather datasets.