OpenAI expanded its Daybreak cybersecurity defense program on August 10, 2026, adding a new AI model, two service tiers, and partnerships with 16 major cybersecurity vendors, according to CyberScoop.
The expansion introduces two tiers, Daybreak Red and Daybreak Blue. Daybreak Red is built for authorized vulnerability research, security testing, and exploit validation, and gives users access to a new cybersecurity-specific model called GPT-5.6-Cyber, OpenAI News said. Partners who receive Daybreak approval can access frontier cyber models and use them to deliver governed cybersecurity services to their own customers.
On August 11, 2026, OpenAI and AWS said Daybreak capabilities are also available through Amazon Bedrock, a move designed to support enterprise security workflows.
TechCrunch reported that AI-led cyberattacks have been increasing, a trend that frames the program's expansion.