MODEL Google/DeepMind Last updated:

GraphCast

AI Weather Forecasting at ECMWF Quality

DeepMind's November 2023 weather model — a graph neural network trained on 39 years of ERA5 reanalysis data, producing 10-day global weather forecasts in ~1 minute on a single TPU. Beats ECMWF's gold-standard physics-based forecast on most metrics. Published in Science. Inspired the entire AI-for-weather field.

Cost
Free
Open weights — self-host
How are Intelligence, Speed & Cost bucketed?
Intelligence and Speed buckets are percentile ranks on Artificial Analysis. Cost buckets are fixed dollar thresholds keyed off output-token price ($/M out).
Intelligence
  • Top 1%≤ 1%
  • Top 5%≤ 5%
  • Top 10%≤ 10%
  • Good≤ 25%
  • Medium≤ 50%
  • Below avg> 50%
Speed
  • Top 1%≥ 345 tok/s
  • Top 5%≥ 237 tok/s
  • Top 10%≥ 196 tok/s
  • Good≥ 146 tok/s
  • Medium≥ 90 tok/s
  • Slow< 90 tok/s
Cost
  • Freeopen weights · self-host
  • Low< $1 / M out
  • Moderate$1–5 / M out
  • High≥ $5 / M out

Why it matters

First credible "AI replaces an HPC pipeline" demonstration in weather. Spawned NVIDIA FourCastNet, Microsoft Aurora, Huawei Pangu-Weather, Shanghai AI Lab FengWu, Fudan FuXi, IBM+NASA Prithvi-WxC, Google MetNet, and ECMWF's operational AIFS — a complete AI-weather industry in 2 years.

Core Capabilities

Science
Built for biology, chemistry, materials, weather, or math research.
Generative
Produces images, video, audio, or other media.
Vision
Understands images, scenes, and visual context.

Context Window

Context window not disclosed.

Availability

API
Not available
Product / App
Not available
Open Source
Released
Enterprise

Pricing Model

Free / self-host
Open weights — pay only for compute.
Self-host

What it feels like

  • Vision-language model from DeepMind — see the linked sources below for benchmark and review coverage
  • Vision and multimodal tasks are the typical fit per the published model card

Best use cases

  • Vision tasks (charts, documents, images) per the model card
  • See the model spec and sources block for benchmarked use cases

Tools to try

Not ideal for

  • Tasks far outside the modalities listed in this model's spec
  • Workflows where a more recent successor in the same family scores higher

Model Evolution

View full evolution tree →

Lam, R. · Sanchez-Gonzalez, A. · Willson, M. · et al.