WeatherNext 3: Making Global Weather Predictions with 5 km Accuracy
Google is adopting a DeepMind AI weather model with 5 km resolution, 60% more accurate precipitation predictions, and a real-time forecast that outperformed all others in Operational WeatherBench.
The world is following biology and chemistry. In meteorology, AI will drive accuracy. Google has just added WeatherNext 3 to Google Search, Maps, and Gemini. It will be the default way hundreds of millions of users get their forecast.
WeatherNext 3 is a new AI weather model developed by DeepMind. Compared to the previous version, WeatherNext 3 boasts a resolution of 5 km, instead of 25 km, improves precipitation prediction accuracy by 60%, provides hourly predictions (instead of every six hours), and is the first one to be deployed in real-time on satellite data. Independently evaluated by Operational WeatherBench, WeatherNext 3 performed better than competing models developed by Microsoft, Nvidia, European Centre for Medium-Range Weather Forecasts, and the National Weather Service of the United States.
Source: techcrunch.com
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →
Author
Evgenii Arsentev
PhD · Chief Executive Officer, digital health
Articles · Latest articles