GenCast is published with a like-for-like comparison against an operational ensemble
DeepMind published GenCast in Nature, reporting that it outperformed the ECMWF ensemble on 97.2% of 1,320 evaluation targets while producing a fifteen-day global forecast in about eight minutes on a single accelerator, against hours of supercomputer time. The comparison was against an operational system's own output on that system's own measures, which is a far stronger evidential position than a self-selected benchmark.
Historical — it happened, and the record is settled.
Subjects
- Machine-learned weather forecasting · technology system · not located
- Google DeepMind · company · United Kingdom · not located
Evidence
Verified — Every field is supported by a cited authoritative source.
ECMWF's AI forecasts become operational, with the GenCast evaluation
ECMWF's Artificial Intelligence Forecasting System became operational on 25 February 2025, outperforming the physics-based system on many measures including tropical cyclone tracks with gains of up to 20%; GenCast, published in Nature in December 2024, is reported to have beaten the ECMWF ensemble on 97.2% of 1,320 evaluation targets and to produce a fifteen-day global forecast in about eight minutes on a single accelerator. V55 verification basis: neither ecmwf.int nor nature.com was fetched. The operational date and the up-to-20% figure were carried in search results attributing them to ECMWF's own news item; the 97.2%, 1,320-target and eight-minute figures were carried with attribution to the Nature paper and to reporting of it.
European Centre for Medium-Range Weather Forecasts; Nature (Springer Nature) · 2024-12 to 2025-02 · Verified
ECMWF news, "ECMWF's AI forecasts become operational", 25 February 2025, https://www.ecmwf.int/en/about/media-centre/news/2025/ecmwfs-ai-forecasts-become-operational ; Price, I. et al., "Probabilistic weather forecasting with machine learning", Nature (December 2024), https://www.nature.com/articles/s41586-024-08252-9
Recorded on Machine-learned weather forecasting
- AlphaGo defeats Lee Sedol in Seoul
- AlphaFold2 is assessed blind at CASP14
- GNoME is announced as 2.2 million new crystal structures
- GenCast is published with a like-for-like comparison against an operational ensemble
- ECMWF makes a machine-learned forecast model operational
Related
Related because they share a subject in the Atlas — never because the text looks similar.
- ECMWF makes a machine-learned forecast model operational · 2025-02-25
- GNoME is announced as 2.2 million new crystal structures · 2023-11
- AlphaFold2 is assessed blind at CASP14 · 2020 (CASP14)
- AlphaGo defeats Lee Sedol in Seoul · 2016-03-09 to 2016-03-15