OpenAI has expanded availability of its purpose-trained cybersecurity models to U.S. government customers through AWS GovCloud (US-East). Daybreak Red, which provides access to GPT-5.6 Cyber on Amazon Bedrock, is now available to approved users, building on the earlier availability of Daybreak models on Bedrock. The offering is designed for authorized vulnerability research, exploit validation, and security testing, giving government security teams a way to apply advanced cyber capabilities within the AWS environments they use for critical missions.

For government organizations, adopting such capabilities involves more than model performance; they must also operate within established security, governance, and operational requirements. Availability in AWS GovCloud (US-East) is positioned as a path for eligible teams to incorporate GPT-5.6 Cyber into those workflows through Amazon Bedrock. Defenders can use the model to reason through security findings, assess impact, and evaluate fixes, helping connect vulnerability research to practical defensive improvements.

Daybreak Red provides verified defenders greater latitude for authorized cybersecurity work, with fewer refusals for these tasks alongside stronger identity verification, monitoring, and access controls. OpenAI notes that for Daybreak Red on Amazon Bedrock, inference data is not used to train models, and customers do not need to share their Daybreak Red data with OpenAI.

Access requires both OpenAI Daybreak Red approval and AWS provisioning. Interested parties must complete the Daybreak application or contact their OpenAI or AWS account team about eligibility. After approval, they must work with their AWS account team to enable access; Daybreak Red approval does not automatically enable model access on AWS. The source also advises qualified security practitioners to review findings and proposed fixes before implementation.