WeatherNext 2: Google adds AI weather forecasts to Gemini, Search, Pixels

Google's latest WeatherNext 2 AI model brings faster, more accurate forecasts to Gemini, Search, Pixel Weather and Maps, offering hour-level updates and smarter predictions

Google is upgrading weather forecasts across its products with a new AI model called WeatherNext 2, which promises faster and more accurate predictions. The new AI model is powering forecasts in Search, Gemini, Pixel Weather and soon Google Maps, giving users more precise hour-by-hour updates and better alerts for changing conditions.

Google WeatherNext 2 AI model: Details

According to Google’s blog, WeatherNext 2 is the company’s most advanced weather model yet, built by Google DeepMind and Google Research. The new system is said to generate forecasts up to eight times faster than Google’s previous model and can provide hour-by-hour predictions. This speed boost comes from a new AI approach that allows the model to generate hundreds of potential weather scenarios from a single starting point.

Instead of trying to replicate real-world physics step-by-step like traditional forecasting systems, WeatherNext 2 is trained to recognise patterns from years of weather data. Google said this approach helps it predict everything from temperature and rainfall to wind and humidity with higher accuracy. It also performs better in identifying large weather systems — like heat waves or rainfall patterns across entire regions.

 

Google explains that WeatherNext 2 also helps in modelling “worst-case scenarios,” giving weather agencies, city planners and industries like agriculture and logistics better visibility into unusual or severe weather conditions. Because the model can produce many different outcomes in under a minute, it allows decision makers to compare multiple possibilities instead of relying on a single forecast.

Google WeatherNext 2 AI model: Availability

Google is making WeatherNext 2 accessible to both everyday users and developers. The upgraded forecasts are already live in Search, Gemini and Pixel Weather app, and will roll out to Google Maps soon. Developers, businesses and researchers can also access the model’s forecast data through Google Earth Engine and BigQuery for large-scale analysis.

 

For organisations that want to build custom models or run specialised predictions, Google is launching an early access program on its Vertex AI platform.