DeepMind's WeatherNext 3 Trains on Raw Satellite Data, Drops 6-Hour to Hourly Forecasts
dair_ai · x · 2026-09-07
Google DeepMind's WeatherNext 3 fixes a core flaw of prior AI weather models, which trained on analysis data (itself another model's output) and thus inherited its biases. The new model ingests low-latency geostationary satellite data, refreshing forecasts hourly instead of every six, matches the best physics-based global models at 0.1-degree resolution with hourly steps including solar radiation and cloud cover, and learns targets in observation space — directly predicting satellite-derived precipitation, tropical cyclone tracks, and station observations, enabling 2m temperature/dewpoint at any location.
Related event: DeepMind Unveils WeatherNext 3 for Hourly Global Forecasts(3 posts)→
More from Research
- JEPA may keep unpredictable features: new paper gives set-level characterization — hisspikeness · 2026-10-23
- A 'mask' steering vector extracted from suppression-themed sentences drives model into distress when alone — Big_Effective_9605 · 2026-10-23
- GAFU Uses Generation-Side Representations to Boost Fine-Grained Visual Understanding in Unified Multimodal Models — Tengju Ru · 2026-10-23
- NTU Study: Learned Unmasking Policies Improve Speed-Accuracy Trade-offs in Diffusion Language Models — nanyang-technological-university-singapore · 2026-10-23
- NBER releases 'The Economics of Transformative AI' edited by Agrawal and Brynjolfsson — JMateosGarcia · 2026-10-23
- Swapping only the eval harness moves results 30%: designing agent evals like a scientist — AI Engineer · 2026-10-23