Here is today's AI news. Story 1. OpenAI. OpenAI launched GPT-6 Astra on September 3, describing it as a major frontier-model upgrade focused on professional workflows, coding, cybersecurity, and browser use.. OpenAI began a limited rollout of GPT‑6 Astra on Sept. 3, calling it its most capable broadly deployed model. Built for coding, research, browser/computer use and document work, it is OpenAI’s first model rated “Critical” for cybersecurity capability. The gated launch underscores how frontier agents may boost professional productivity while requiring stricter access and safety monitoring. Story 2. Nvidia. Nvidia agreed to acquire AI platform Hugging Face for approximately $13 billion, expanding its position in open-model infrastructure and developer tools.. Nvidia agreed to buy Hugging Face for $12.93 billion, including up to $1 billion in employee-retention equity. The deal gives Nvidia direct reach to a major open-model developer platform used by 18 million people and 200,000 companies. Jensen Huang said Hugging Face will remain “an open platform,” with no requirement to use Nvidia chips. Story 3. Google DeepMind. Google introduced WeatherNext 3, an advanced AI weather model designed to improve global forecasting accuracy and speed.. Google DeepMind and Google Research introduced WeatherNext 3, an AI forecasting model that ingests live geostationary-satellite data and refreshes predictions hourly. It delivers forecasts up to 15 days ahead, with resolution as fine as 5 km for selected surface variables—about five times sharper than WeatherNext 2. Google reports up to 50% lower precipitation error. The model now supports Search, Gemini, Maps and clean-energy planning. Story 4. OpenAI. OpenAI committed $1 billion in subsidized cybersecurity tools, training, and technical support for organizations protecting critical infrastructure from AI-enabled attacks.. OpenAI announced “Daybreak for Frontline Defenders,” a $1 billion commitment in subsidized cybersecurity-model access, training and technical support, targeted for use over six months. Initial U.S. beneficiaries include water and wastewater utilities, electric-grid operators, state and local governments, community banks, nonprofits and open-source maintainers. The tools can review legacy code, investigate suspicious activity, validate vulnerabilities and test fixes. An MS-ISAC pilot will train public-sector and water-system defenders, expanding AI-powered protection for resource-constrained critical infrastructure. Story 5. Microsoft. Microsoft began offering OpenAI’s GPT-6 Astra through the Foundry Limited Access Program, bringing the new model to selected enterprise customers.. Microsoft says OpenAI’s GPT‑6 Astra is rolling out via the Microsoft Foundry Limited Access Program to approved enterprises. The staged launch reflects tighter controls for frontier models with advanced cyber and agent capabilities. Azure documentation lists a 1.05M-token context window, image/text input, tool calling and multi-agent orchestration; OpenAI calls it its “most capable” broadly deployed model. That's all for today's AI news briefing.