Google's WeatherNext 3 refreshes its forecast every hour
Sasha / Models and Research desk
Google’s weather AI now updates every hour, grounded in the latest satellite pass.
What happened
Google DeepMind launched WeatherNext 3, which it calls the first global weather model to generate hourly forecasts. Rather than relying only on physics simulations, it ingests a mosaic of live geostationary satellite data and produces forecasts at up to 5km resolution, compared with the 25km grid of the previous version, a picture roughly five times sharper. Google reports precipitation accuracy gains of up to 60 percent on medium-range forecasts against one benchmark. The model is being integrated into Search, Maps, the Gemini app and Google’s developer tools.
Why it matters
Hourly, satellite-grounded forecasts at neighborhood scale are the kind of quiet infrastructure upgrade that reaches billions of people through products they already use, and a further sign of AI weather models moving from research to daily service.
Sources
Related
GPT-6 Astra drew Michael Jackson in Google Calendar
A viral demo shows GPT-6 Astra recreating images inside Google Calendar by placing colored events across the week grid, block by block: Michael Jackson, a cat, fictional characters. The output is crude, but the mechanism is the point: the agent operates an ordinary app the way a person would, watching the screen and clicking, toward an entirely decorative goal.
Gemini 3.8 Flash is out, Google's third Flash release in six weeks
Google released Gemini 3.8 Flash on September 2, 2026, calling it "our most intelligent workhorse model" and noting three Flash versions in six weeks. It scores 54.9% on HLE-Verified and outperforms most larger frontier models on the DeepSWE v1.1 coding benchmark. Pricing is $0.75 per million input tokens and $3.75 output through December 31, 2026, then $1.50 and $7.50 from January 1, 2027. A Gemini 3.8 Flash Cyber variant goes to trusted testers through the Fairwind Program.
An attacker used AI agents to build a credential harvest in under six hours
Google Threat Intelligence Group described an incident in which a suspected financially motivated attacker compromised a cloud resource, then planned, built and executed a mass credential harvesting campaign in under six hours. The attacker assembled an autonomous framework from an AI coding chatbot, a prompt and a set of markdown instruction playbooks that drove automated scanning and harvesting. The system produced a dashboard that organized and validated more than 23,800 harvested secrets in real time, including API keys for cloud and AI services. Google observed the campaign in the second quarter of 2026.
ANOTHER News is published by ANOTHER, an AI-native content agency. Daily coverage also runs on Instagram.