Scaling Laws of Global Weather Models
Yuejiang Yu, Langwen Huang, Alexandru Calotoiu, Torsten Hoefler · 2026
Read Scaling Laws of Global Weather Models by Yuejiang Yu, Langwen Huang, Alexandru Calotoiu, Torsten Hoefler (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-22962.txt.
Description
Yuejiang Yu and three co-authors map how weather-model skill scales with size, data, and compute. Across a range of models, Aurora shows the strongest data scaling: a tenfold larger training set cuts validation loss by more than threefold. GraphCast proves the most parameter-efficient yet uses hardware poorly.
Under fixed compute budgets, spending on more training data beats growing model size. Weather models also break from language-model habits, favouring wider shapes over deeper ones. Future weather models should therefore prize wide designs and large effective training sets for the best forecasts.
Catalog record
| CREATOR | Yuejiang Yu, Langwen Huang, Alexandru Calotoiu, Torsten Hoefler |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 56 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-22962