OpenAI Daybreak: $1B Frontline Defenders Program, Blue vs Red and AWS Access
OpenAI expanded its Daybreak cyber-defense program on September 3, 2026 with a $1 billion global commitment for subsidized model access, training, technical support and partnerships. The company says the subsidy is targeted for use over the next six months, beginning with resource-constrained defenders in the United States and expanding to partner countries.
Priority groups include water and wastewater operators, electric-grid operators, state and local governments, community and regional banks, nonprofits and open-source maintainers. OpenAI also announced a public-sector and water-system pilot with the Multi-State Information Sharing and Analysis Center (MS-ISAC) and says the Daybreak Defense Network now includes more than 35 enterprise products and partner-operated services.
Daybreak remains a governed-access cybersecurity platform. Daybreak Blue is the general starting tier for verified defenders, while Daybreak Red provides more permissive access for advanced authorized vulnerability research, exploit reproduction and mitigation development. Eligible customers can also use both tiers through Amazon Bedrock.
September 3 expansion: the key facts
| Item | Current detail |
|---|---|
| New commitment | $1 billion in subsidized Daybreak access, training, support and partnerships |
| Target consumption window | Six months |
| Initial geographic focus | United States, followed by partner countries |
| Priority groups | Critical infrastructure, state/local government, community banks, nonprofits, open-source maintainers and other resource-constrained defenders |
| Public-sector pilot | MS-ISAC, focused initially on public-sector and water-system defenders |
| Existing Daybreak adoption | Thousands of defenders across 2,000 approved organizations and workspaces, according to OpenAI |
| Partner ecosystem | More than 35 enterprise products and partner-operated services through the Daybreak Defense Network |
OpenAI says participating organizations can use Daybreak to review legacy code and configurations, analyze suspicious activity, identify and validate vulnerabilities, prioritize remediation, and develop and test fixes.
The company says it previously offered affected U.S. states and utilities up to $1 million in API credits, Daybreak access and technical assistance following attacks on water systems. Those deployments were used for code and configuration review, validation, patch development and verification while services remained operational.
Daybreak Blue vs Red
| Area | Daybreak Blue | Daybreak Red |
|---|---|---|
| Primary model path | GPT-5.6 Sol | GPT-5.6-Cyber |
| Intended work | Vulnerability discovery, secure-code review, detection engineering, malware analysis, incident response and patch validation | Advanced vulnerability research, exploit reproduction, exploit validation and mitigation development |
| Typical starting point | Most approved defensive teams | Vetted teams requiring higher-risk cyber capability |
| Access model | Verified Daybreak access | Higher-risk approval with stronger identity, monitoring and scope controls |
OpenAI describes Daybreak as a governed cyber-defense stack combining frontier models, the Codex harness, Codex Security, trusted workflows and ecosystem partners. Its current workflow model covers inventory, discovery, dynamic validation, ownership assignment, verified remediation and repeated verification.
Who can apply for the $1B Frontline Defenders program
OpenAI's September 3 announcement prioritizes organizations that protect essential services while operating with limited security resources. The stated priority population includes:
- water and wastewater systems;
- electric-grid operators;
- state and local governments;
- community and regional banks;
- nonprofits;
- open-source maintainers;
- other organizations responsible for essential digital infrastructure.
OpenAI says eligible state and local governments, critical-infrastructure operators, nonprofits, open-source maintainers and supporting organizations can register interest through the Daybreak program.
The initiative also adds hands-on assistance. OpenAI says a recent utility-focused gathering included participants representing 40 states and the District of Columbia, collectively serving more than half of the U.S. population.
MS-ISAC pilot
The new MS-ISAC pilot is aimed at state, local, tribal and territorial cyber defenders, beginning with public-sector and water-system participants.
The program combines Daybreak access with guided training and technical assistance for tasks such as validating findings, prioritizing remediation and developing repeatable defensive workflows. MS-ISAC already provides threat intelligence, incident-response support and shared defensive services to thousands of public-sector organizations.
For smaller public-sector and utility teams, the practical value is the combination of subsidized compute access and specialist support. Model access alone has limited operational value when organizations lack the staff or processes required to validate findings and move fixes into production safely.
Daybreak on Amazon Bedrock
AWS made Daybreak Blue and Daybreak Red available to eligible Amazon Bedrock customers in August 2026. AWS currently documents:
| Tier | Bedrock model | Context window | Supported use |
|---|---|---|---|
| Daybreak Blue | GPT-5.6 Sol | 1M tokens | Verified defensive cybersecurity work |
| Daybreak Red | GPT-5.6 Cyber | 272K tokens | Advanced vulnerability research, exploit reproduction and mitigation development |
AWS lists both models as active and requires prior Daybreak eligibility. Its launch documentation places the service in US East (Ohio) for eligible customers.
AWS says both models run on its next-generation Bedrock inference engine with zero-operator access enforced at the chip. AWS also states that inference data is excluded from model training and customers are not required to opt into sharing inference data with OpenAI.
GPT-5.6-Cyber and the higher-risk tier
GPT-5.6-Cyber is the specialized model associated with Daybreak Red. It is designed for advanced authorized cybersecurity tasks where standard model safeguards can interrupt legitimate research workflows.
OpenAI's earlier Daybreak evaluations reported a 95% Advanced Cybersecurity Completion Rate for GPT-5.6-Cyber in Daybreak Red. That internal metric measures whether the model completes advanced cyber requests, including exploit-chain development, authentication bypass and privilege escalation tasks. The metric measures workflow completion in OpenAI's evaluation; it does not measure compromise success against real-world targets.
OpenAI also reported that GPT-5.6-Cyber helped researchers identify previously unknown V8 vulnerabilities, including one assigned CVE-2026-15903, as part of coordinated disclosure to Google.
Daybreak expands its defense delivery program
The September 3 expansion adds subsidized access, training, technical assistance and partner-operated services to Daybreak's controlled model-access program. Deployment pathways now include OpenAI and eligible Amazon Bedrock accounts.
Resource-constrained defenders still need asset context, authorization boundaries, validation, ownership, patch development, testing and verification to operationalize model-generated findings safely.
OpenAI's current Daybreak architecture reflects that operational sequence. The platform is being positioned around a continuous defensive loop in which findings move from discovery through validation and remediation while consequential actions remain subject to human review and governance controls.
Bottom line
The $1 billion Frontline Defenders commitment materially expands Daybreak's reach through a six-month subsidy window, an MS-ISAC pilot, training and technical support, and more than 35 partner products and services.
Organizations evaluating Daybreak should start with three questions: whether they qualify for subsidized access, whether Blue provides sufficient capability for their workflow, and whether their security processes can validate and safely operationalize model-generated findings. Red is suited to approved teams that specifically require advanced exploit research and mitigation-development capability.
Sources
- OpenAI, Daybreak for Frontline Defenders: $1B to protect essential services, September 3, 2026: https://openai.com/index/daybreak-for-frontline-defenders/
- OpenAI, Daybreak: https://openai.com/daybreak/
- AWS, Daybreak Red and Daybreak Blue from OpenAI are now available to eligible customers on Amazon Bedrock, August 13, 2026: https://aws.amazon.com/about-aws/whats-new/2026/08/openai-daybreak-red-and-blue-on-amazon-bedrock/
- AWS, Daybreak Red: GPT-5.6 Cyber model card: https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-56-cyber.html
- AWS, Daybreak Blue: GPT-5.6 Sol model card: https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-openai-gpt-daybreak-blue-56-sol.html