OpenAI has expanded the distribution of its Daybreak cybersecurity models. The company says enterprises can now access Daybreak capabilities through Amazon Bedrock to support security workflows. This is the latest move to put its cyber AI in front of enterprise customers.
OpenAI also introduced GPT-5.6-Cyber, a model built specifically for cybersecurity work. It is available through Daybreak Red for authorized vulnerability research, exploit validation, and security testing. The company further said approved Daybreak partners can deliver governed cybersecurity services using frontier models, an approach meant to keep powerful capabilities in trusted hands.
The announcements also cover safety and evaluation. OpenAI released preliminary cybersecurity evaluations for Astra and described steps to strengthen safeguards. In a separate update, it addressed recent third-party cyber evaluation incidents and outlined new safeguards for model testing. The sources describe different components of the same expansion, but they converge on a common narrative: broader access to frontier cyber models, paired with stricter oversight and authorization.