Skip to content

NeuralGCM

A climate model that mixes physics and machine learning

Visit official site
Category AI models

What NeuralGCM is

The model is described as a climate model that combines traditional physics with machine learning to improve the accuracy and efficiency of simulations.

What you can do with it

  • Generate 2 to 15 day weather forecasts
  • Reproduce 40 years of temperature data
  • Learn small-scale physics from existing weather data

Application scenarios

  • Climate scientists forecast trends for a region
  • Government agencies plan for extreme weather events
  • Teaching institutions explain the climate system

Getting started

  • Download the source code and model weights
  • Install the required dependencies
  • Feed in a weather dataset and set resolution and time range

Pros & cons

✓ What we like

  • Runs efficiently on TPUs and GPUs
  • Gradient-based tuning of the solver is supported

! What to watch out for

  • The model is stated to be not yet a full climate model
  • The workflow needs weather datasets as input

FAQ

How long are the forecasts?

The feature list states 2 to 15 day weather forecasts.

What does it run on?

The material says it runs on TPUs and GPUs, unlike models that mainly use CPUs.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review