OpenAI is tightening controls around Astra after cybersecurity evaluations suggested the upcoming AI model may be approaching the company’s Critical capability threshold. Astra has not been formally classified as Critical, but OpenAI said recent tests showed major advances in agentic coding and cyber abilities. Some internal work that does not meet the new security requirements is being paused while testing continues under stricter safeguards. The company is also expanding its Daybreak cybersecurity program and introducing GPT-5.6-Cyber for approved security research. For now, Astra remains under evaluation as OpenAI studies how powerful its cyber capabilities may have become.
by rrollins
Recent Comments