Google launches WeatherNext 3, its most accurate AI weather model, trained on real satellite data
geoffwolfe · x · 2026-09-04
Google DeepMind and Google Research introduced WeatherNext 3, their most advanced global AI weather model, generating hourly high-resolution forecasts, precise precipitation predictions, and clean-energy variables.
The key differentiator: most AI weather models train on NWP model outputs with a six-hour lag, but WeatherNext 3 trains on real-world satellite observations, bypassing that constraint. It now powers forecasts in Search, the Gemini app, and Google Maps, and joins Brightband's Operational WeatherBench leaderboard.
Related event: Google Launches WeatherNext 3, Its Most Advanced Global Weather AI Model(19 posts)→
More from Models
- Opus 5.5 tops OpenRouter spending in 20 of 29 task categories — mmurph · 2026-10-21
- Meta team builds open-source 30B LLM from scratch in months, hails 'second era' — TimDarcet · 2026-10-21
- DFlash drafter for Muse Glimmer-30B: 5 layers, blockwise 16-token spec decoding — TimDarcet · 2026-10-21
- EvoPress quantization shrinks 30B model from 56GB to 19GB at near-bf16 accuracy — TimDarcet · 2026-10-21
- Radar chart: Muse Glimmer-30B benchmarked against Gemma4-31B and Qwen3.6-27B — TimDarcet · 2026-10-21
- Privacy evals: Muse Glimmer-30B leaks less than Qwen3.6-27B on CIMemories — TimDarcet · 2026-10-21