Google's WeatherNext 3 reads raw satellite data, cuts short-range temperature error up to 40%

anselm · x · 2026-09-24

Google DeepMind and Google Research published the paper behind WeatherNext 3, tackling two key flaws of AI weather models.

• Escaping analysis-data lag: Existing AI models start from analysis fields updated every 6 hours, inheriting 6-12 hour delays and their biases. WeatherNext 3 ingests raw observations instead, building an hourly global mosaic from geostationary satellite feeds with 1 hour latency, issuing forecasts every hour.
• Physics-model-grade resolution: Hourly time steps and 0.1-degree resolution for single-level variables including solar radiation and cloud cover.
• Results: Up to 40% lower short-range temperature error at stations never seen in training, and a new state of the art in probabilistic medium-range forecasting skill.

Original post →

More from Research

• Stanford's DyCraft Uses Agentic Decomposition to Cut Physical Errors in Long-Horizon 4D Scene Generation — StanfordUniversity · 2026-11-11
• Bert Chan on His Unconventional Path from Side-Projects to the ALife Institute — BertChakovsky · 2026-11-11
• CGM Study of 62,000 People: Average Glucose Beats HgbA1C at Detecting High Insulin — arjunrajlab · 2026-11-11
• First open dictation benchmark: 761 recordings across 10 apps, Wispr Flow 79% vs Apple 35% — rohanpaul_ai · 2026-11-11
• Boltz models showcased at MSK Cancer Center for accelerating early-stage drug discovery — charlieharris01 · 2026-11-11
• Carlman: a Pac-Man-style game where all characters are Lenia cellular-automata solitons — BertChakovsky · 2026-11-11

Research channel →