Google DeepMind and Google Research have introduced WeatherNext 3, described as the most advanced and accurate global weather model to date according to independent live evaluations by Brightband. The flagship AI forecasting model incorporates real-time satellite data, hourly refreshes, higher resolution, precise precipitation forecasting, and new clean energy variables. It is now integrated across Google Search, the Gemini app, Google Maps, Google Maps Platform, and Google Cloud.

Every day, the weather influences billions of decisions. Some are as simple as grabbing an umbrella before heading out the door, but others are far more consequential. Wind, rain, and extreme weather events such as heatwaves and droughts have cascading impacts across agriculture, global supply chains, clean energy production, and national economies.

In recent years, AI has revolutionized weather forecasting, using historical records to make faster and more accurate predictions than traditional methods. DeepMind's own GraphCast, released in 2023, outperformed conventional numerical systems on many metrics and helped establish machine learning as a credible complement to physics-based forecasting, alongside efforts from other organizations such as the European Centre for Medium-Range Weather Forecasts (ECMWF) and its open-source AIFS model. Yet predicting highly local and rapidly changing weather has remained a challenge. Previous models often lacked sufficient spatial resolution and struggled to incorporate real-time weather data from sources like satellites.

WeatherNext 3 learns directly from real-time observations, enabling timely and more localized predictions for the weather events that impact people most. By using raw satellite data to produce a high-resolution forecast every hour, the model makes reliable forecasts accessible across Google products worldwide.

Rapid weather prediction at unprecedented resolution

A forecast's utility often comes down to detail — how finely it resolves both time and space. WeatherNext 3 generates hourly forecasts at multiple spatial resolutions, maintaining physical consistency from broad global wind patterns down to local topography.

The model visualizes key surface variables such as temperature and moisture at 5-kilometer resolution, other surface variables at 10 kilometers, and atmospheric variables such as wind speed at 25 kilometers. Overall, this delivers a global weather picture roughly five times sharper than its predecessor, WeatherNext 2, which produced forecasts on a 25-kilometer grid in 6-hour increments.

Architecturally, WeatherNext 3 ingests live 1-hour geostationary satellite mosaics alongside traditional historical analysis to feed a single, flexible Functional Generative Network (FGN) mesh transformer. It outputs dense gridded fields, discrete cyclone tracks, and station-level sparse predictions natively.

In a comparison of 2-meter temperature forecasts over the UK, WeatherNext 3's native 5-kilometer (0.05°) resolution resolves intricate local topography, avoiding the pixelated, over-smoothed thermal representations produced by WeatherNext 2 at 25-kilometer (0.25°) resolution.

Real-world data at continuous global scale

The biggest leap forward is what the model learns from. Most AI weather models, including WeatherNext 2, are trained on data from numerical weather prediction (NWP) models. Although useful, NWP models are complex, supercomputer-driven physics simulations that carry a six-hour data lag — a delay that can introduce biases for fast-changing variables like rain or surface temperature.

By ingesting a mosaic of live, global geostationary satellite data, WeatherNext 3 gains a continuously updating view of the atmosphere. This allows the model to generate a new forecast every hour, each grounded in the most recent satellite observations available, at up to 5-kilometer resolution.

This rapid refresh matters because critical weather develops quickly. When storms, fronts, or precipitation systems materialize suddenly, the model's hourly update cycle and higher resolution provide earlier, more detailed insights to support effective response.

Some variables, like temperature and humidity, can fluctuate dramatically over just a few kilometers — particularly relevant for communities near coastlines, valleys, or mountain ranges. Traditional models struggle here because they train on atmospheric representations that lack detail and miss extreme local variations. WeatherNext 3 instead trains directly on sparse weather station observation data, allowing global forecasts on a 5-kilometer grid that account for regional details like topography.

According to Google, this breakthrough is especially significant for regions across Latin America, Africa, and Asia-Pacific that have historically been underserved by high-resolution forecasting due to the immense supercomputing costs of traditional regional models. It brings localized, high-fidelity forecasting to billions of people and local businesses in these areas. Because AI models run on far less compute than supercomputer-based NWP systems once trained, they lower the cost barrier that has historically limited high-resolution forecasting to wealthier regions.

Beyond resolution and forecast frequency, the model introduces predictions engineered for renewable energy production. WeatherNext 3 forecasts 100-meter wind speeds — roughly at turbine height — for precise wind-energy output estimation, alongside high-resolution cloud cover and solar radiation levels to help solar farms estimate how much light will reach the ground. Google says this data is crucial for global clean energy planning, allowing grid operators and renewables developers to predict how much power their clean energy assets will generate and match it with consumer demand. That capability aligns with Google's own stated goal of operating on carbon-free energy around the clock by 2030, an effort that depends heavily on predicting renewable output.

Precipitation forecasting at breakthrough accuracy

Global weather models have long struggled to predict precipitation accurately. Rain and snow systems are driven by fast-moving cloud processes on tiny scales that are difficult to model with traditional physics-based simulations. As a result, AI forecasts often produce blurry estimates or miss the boundaries of severe storms entirely.

To address this, Google trained the model on two high-quality precipitation datasets: NASA's satellite-based Integrated Multi-satellite Retrievals for GPM (IMERG) and its own global precipitation reanalysis based on satellite radar.

The result is a significant leap in precipitation forecasting accuracy. In medium-range global forecasts, evaluations against baselines show a Continuous Ranked Probability Score (CRPS) improvement of up to 60% against IMERG, 30% for MRMS, and 10% against rain gauge measurements for early lead times.

In a medium-range probability-of-precipitation (PoP > 1mm) forecast comparison, WeatherNext 2 at 25-kilometer resolution showed a highly diffused and pixelated precipitation footprint, while WeatherNext 3 at 11-kilometer resolution closely mirrored actual satellite ground truth, accurately capturing the sharp, convective bands of the weather systems.

Research applied across the ecosystem

Google states that its primary goal is to advance weather intelligence to make it universally useful — whether for an emergency responder tracking sudden wind shifts, an air traffic controller planning flight paths, or a farmer managing crops.

To bring these breakthroughs out of the lab, WeatherNext 3 is being integrated across Google's core ecosystem and beyond:

• High-resolution forecast data: Global weather predictions are updated hourly and ready to integrate into user workflows with no model setup required. Researchers, developers, and businesses can query the data in BigQuery and Earth Engine, or bulk-download it from Google Cloud Storage.
• Global availability: WeatherNext 3 begins powering weather experiences within Google Search, the Gemini app, Google Maps, Google Maps Platform Weather API, and Google Earth Engine starting today. The update dramatically improves longer-term forecasts: when planning a day or more ahead, users will see up to 50% more accurate precipitation forecasts, with the greatest improvements in regions where forecasts have historically been less reliable.

Google notes that the atmosphere will always retain a degree of unpredictability. However, by training on real-world observations and bypassing traditional modeling constraints, WeatherNext 3 brings forecasts closer to matching what is actually happening on the ground. As AI forecasting matures, a recurring question in the field is how machine-learning systems and traditional physics-based agencies will coexist — national weather services, including those collaborating with AI developers, continue to serve as the authoritative source for warnings and public safety, a role Google itself acknowledges.

Additional resources related to the release include the WeatherNext 3 research paper, developer documentation for building with WeatherNext 3, the Weather Lab for real-time visualizations of the model, and Brightband's independent live leaderboards showing where WeatherNext 3 ranks. Readers can also explore Google's broader geospatial and AI work through Google Earth Engine, AlphaEarth Foundations, and Earth AI.

Google's disclaimer notes that for official weather forecasts, severe weather warnings, and public safety advisories, users should refer to their local meteorological agency or national weather service.

Source: Google DeepMind Blog