GPT-6 Astra is OpenAI’s newest frontier AI model, designed to move beyond answering questions toward carrying out complex, multi-step work using computers and software. Astra combines advanced reasoning with computer use, web research, coding, scientific analysis, and the ability to create polished documents, spreadsheets, presentations, websites, and other professional work products.
Astra can remain oriented during long workflows, adapt when instructions change, and make routine decisions while asking users about consequential choices. Astra is also showing considerable promise for scientific discovery. At the same time, its increased capabilities—particularly in cybersecurity and autonomous computer use—create new safety challenges, leading OpenAI to deploy additional monitoring and safeguards intended to keep the model within authorized boundaries.
GPT-6 Astra represents an important step toward agentic AI: systems that do not simply tell people how to accomplish a task, but can increasingly plan, use digital tools, evaluate their own progress, and carry substantial portions of the work through to completion.
The video above is intended for educational purposes and does not constitute or imply an endorsement by FutureForAll.org
