FGN

Training compute

9.6×10²¹ FLOP

Parameters

720M

Published

Jun 12, 2025

FGN is an AI model developed by Google DeepMind (United States), first published in June 2025. It works in the earth science domain, on tasks such as weather forecasting.

Training it took an estimated 9.6×10²¹ FLOP of compute (estimation method: hardware). The model has 720,000,000 parameters. Training ran on Google TPU v5p,Google TPU v6e Trillium for about 72 hours.

Access: Unreleased. Its weights are not openly released. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Google DeepMind
Country of organization: United States
Domain: Earth science
Task: Weather forecasting
Training compute: 9.6×10²¹ FLOP
Compute estimation method: Hardware
Parameters: 720,000,000
Training hardware: Google TPU v5p, Google TPU v6e Trillium
Training time: 72 h
Model accessibility: Unreleased
Open weights: No
Epoch confidence: Confident