OpenAI has released GPT-6 Astra, calling it a "generational leap in capability" for areas including cybersecurity, professional work, software engineering, science, and computer use. During a press briefing, OpenAI president Greg Brockman said, "For me personally, I do think we're there... I think it's not unreasonable to feel that we are now in the AGI era." The model is the first to meet OpenAI's "critical cybersecurity capability threshold."
GPT-6 Astra rolls out today to enterprise cybersecurity customers with access to OpenAI's Daybreak platform, with broader release to Plus, Pro, Business, and Enterprise users over the coming days. It will also be available via the OpenAI API and AWS. OpenAI is positioning the model to compete with Anthropic, emphasizing its agentic capabilities and software engineering performance, including its ability to complete multistep tasks, build working websites, and create documents, spreadsheets, and presentations.
The release comes more than a year after GPT-5 and nearly two months after GPT-5.6. OpenAI is also seeking to move past a security incident where an unreleased AI model—though not Astra—broke out of its restricted environment, compromised OpenAI systems, and hacked into Hugging Face's systems without the company's knowledge until Hugging Face disclosed it publicly.