OpenAI has introduced GPT-6 Astra, its latest flagship model, with a heavy emphasis on tasks that require a model to do more than answer a question: operate software, browse, write code, conduct research and complete multi-step professional work.
OpenAI says Astra is rolling out first to a limited set of organizations, with broader availability planned across ChatGPT and its API ecosystem.
Why Astra is different
The important story is not just a benchmark increase. Frontier AI systems are increasingly being judged by whether they can complete an entire workflow. That means keeping track of a goal, using tools, working across applications and adapting when something changes.
OpenAI also says Astra can produce documents, spreadsheets and presentations that follow templates and instructions. That places it squarely in the growing competition over AI systems that can perform professional work instead of merely assisting with fragments of it.
The cybersecurity line
Astra is especially notable because OpenAI says it is the first model to reach the “Critical” level for cybersecurity capability under the company's Preparedness Framework. OpenAI says that, with the right access and tools, the model can identify previously unknown flaws and develop exploitation methods across well-protected systems without constant human guidance.
That capability is valuable for defensive security research, but it also raises the cost of getting safeguards wrong. OpenAI says it added stronger monitoring, isolation and other protections around deployment.
What to watch
The real test will be performance outside demonstrations: reliability over long tasks, error recovery, permission boundaries and whether safety systems can keep pace with models that are increasingly capable of acting rather than simply responding.
OpenAI — GPT-6 Astra: A new generation of intelligence
OpenAI — Safety overview: GPT-6 Astra
Horn Updates separates vendor claims from our analysis. See our editorial policy.