Discover the most influential and efficient open-source AI models designed to tackle climate change. This selection highlights tools driving sustainability, environmental monitoring, and weather forecasting. Explore low-carbon machine learning solutions and green AI applications that are transforming climate action. Ideal for developers, researchers, and businesses seeking to implement responsible AI technology for a more sustainable future.
Provides 15-day global forecasts
(+4)
This model provides high-accuracy medium-range weather predictions, outperforming leading open models and offering significant speed advantages. Its Atlas architecture covers over 70 weather variables up to 15 days in advance.
Uses generative AI for cloud evolution prediction
(+4)
Utilizing generative AI, this model predicts the evolution of cloud and rainfall systems with kilometer-resolution forecasts. It's the first AI model to outperform traditional physics-based models for short-term precipitation forecasting.
Enables AI-driven downscaling of climate projections
(+4)
This model uses generative AI to downscale coarse-resolution continental predictions to high-resolution regional weather fields. It provides fine-grain resolution up to 500 times faster than traditional methods.
Produces initial conditions for weather prediction
(+1)
Powered by the HealDA architecture, this model produces initial conditions for weather prediction in seconds on GPUs. It significantly reduces the computational burden of preprocessing raw observational data.
Accelerates AI-based global weather forecasting
(+4)
This model delivers high forecasting accuracy for various weather variables, surpassing leading conventional ensemble models. It produces forecasts up to 60 times faster than traditional methods.
All the rankings you can imagine
Thousands of verified votes to discover the best. Your vote here counts
Predicts hundreds of weather variables globally
(+4)
This established open-source model generates 10-day forecasts at 28-kilometer resolution, outperforming traditional HRES on 90% of test variables. It offers broader hardware compatibility, running on any hardware.
Fast machine learning model for climate simulation
(+4)
This open-source emulator, built on NVIDIA's SFNO architecture, emulates daily weather variability and climate at 100 km resolution. It runs approximately 1600 years/day on a single GPU, making it 100 times faster than comparable physics-based models.
Learns directly from data
(+4)
This new framework for seasonal climate predictions learns directly from data, combining observational records with climate simulations. It uses transformer-based neural networks and variational inference to provide probabilistic predictions.