This week, the DeepMind team at Google has revealed a new artificial intelligence (AI) model for predicting weather, named GenCast.

According to a study featured in Nature, the researchers at DeepMind have discovered that GenCast outshines the ENS of the European Centre for Medium-Range Weather Forecasts, which is currently considered as the globe’s premier operational forecasting system.

DeepMind’s team shared a more user-friendly explanation of the technology in a blog post. In contrast to their previous weather model, which was “deterministic and offered a singular, optimal estimate of future weather,” GenCast “involves a group of 50 or more predictions, each signifying a potential weather trajectory.” This results in a “complicated probability distribution of future weather situations.”

Regarding its performance against ENS, the team stated that they trained GenCast using weather data until 2018, and then analyzed its forecasts for 2019. The results showed that GenCast’s predictions were more precise 97.2 percent of the time.

Google declared that GenCast is a part of its collection of AI-driven weather models, which are gradually being integrated into Google Search and Maps. Additionally, they intend to publish real-time and historical forecasts from GenCast that can be used by anyone for their own research and modeling purposes.