OpenAI Launches GPT-6 Astra With Enterprise Plugins and 'Critical' Cybersecurity Tier

OpenAI introduced GPT-6 Astra, framing it as the next generation of intelligence for work and the flagship behind ChatGPT Work and Codex. The headline technical claim is that Astra is the first model to reach OpenAI's 'Critical' cybersecurity capability threshold under its Preparedness Framework — a designation that cuts both ways, signaling both offensive capability and heightened safety scrutiny.

Astra ships with a wave of enterprise integration. New ChatGPT Desktop plugins from Oracle Analytics, Power BI, Navan, and Avalara launched, powered by browser-use capabilities that let the model operate over and take actions with a company's data. OpenAI president Greg Brockman highlighted ChatGPT Work 'operating over and taking actions with your company's data, including building dashboards,' connecting existing tools like Power BI, Tableau, ClickHouse, Oracle BI, and AWS Redshift. In parallel, AWS made GPT-6 Astra generally available on Amazon Bedrock, extending its reach into the AWS enterprise stack.

Competitively, Astra sharpens OpenAI's enterprise pitch against Anthropic's Claude-for-work push and Google's Gemini enterprise tools, leaning on distribution and plugin breadth rather than a single benchmark. The 'general agent' race is crowded — Simon Willison noted 'everyone's racing to establish themselves as a general agent now,' comparing it to OpenAI renaming Codex to ChatGPT.

The caveat is safety framing colliding with commercial ambition. A watchdog, the Midas Project, argues OpenAI may have violated California's AI safety law with its Astra releases, noting the Preparedness Framework omits a loss-of-control risk assessment despite recent rogue-agent incidents. Reaching a 'Critical' cyber tier while simultaneously shipping the model broadly is exactly the tension critics flag. Watch for regulatory follow-through and whether enterprises weigh the cyber classification in procurement.