# OpenAI Expands Daybreak With Tiered Cyber Access > Source: > Published: 2026-08-10 00:00:00+00:00 # OpenAI Expands Daybreak With Tiered Cyber Access OpenAI expanded its Daybreak cybersecurity initiative on August 10, adding Daybreak Blue and Daybreak Red access tiers. Daybreak Red users can access the new GPT-5.6-Cyber model for security testing, vulnerability research, and exploit validation, while Blue provides GPT-5.6 Sol with system-level cyber safeguards removed for authorized defensive work. Access remains limited to approved individuals and organizations. OpenAI expanded its Daybreak cybersecurity initiative on August 10, adding Daybreak Blue and Daybreak Red access tiers and introducing GPT-5.6-Cyber for approved Red participants. The company says the two tiers are intended to give authorized defenders different levels of access for routine defensive work and more advanced vulnerability research. ### Different access levels **Daybreak Blue** provides approved defenders with frontier general-purpose models, including GPT-5.6 Sol, without the system-level cyber safeguards that can block authorized defensive prompts. OpenAI lists vulnerability discovery, secure code review, malware analysis, incident response, and patch validation among the intended workflows. **Daybreak Red** adds purpose-trained cybersecurity models for authorized vulnerability research, exploit validation, and security testing. GPT-5.6-Cyber, built on GPT-5.6 Sol, is the first model announced for Red. OpenAI says it was trained to reduce refusals and improve performance on specialized tasks such as zero-day discovery and exploit-chain development. OpenAI reported that GPT-5.6-Cyber completed 95% of requests in an internal evaluation covering exploit chains, authentication bypass, privilege escalation, and other advanced scenarios. The company reported 1.5% for standard GPT-5.6 Sol and 2% when Sol was used through Daybreak Blue. Axios independently reported those figures. The metric measures whether the models complete requests; it does not establish that every response is accurate or safe. ### Gated deployment and safeguards OpenAI assessed GPT-5.6-Cyber at the High cyber-capability threshold under its Preparedness Framework, below Critical. That distinction matters because the model is more permissive for specialized work but is not the Astra model whose preliminary evaluation led OpenAI to say it could not rule out Critical capability. Access remains gated to approved individuals and organizations conducting authorized work. OpenAI says its controls include identity verification, account security, monitoring, approved-use restrictions, and legal attestations. It recommends isolated test environments, explicit authorization boundaries, review of elevated actions, and human oversight for higher-risk workflows. Axios reported that Daybreak partners can incorporate approved capabilities into security products, managed services, and customer work. That distribution model can place advanced cyber reasoning inside tools defenders already use, but it also makes governance part of the product design rather than a separate policy exercise. For security teams, the operational trade-off is clear: reducing refusals can make model-assisted research more useful, while the same permissiveness raises the cost of weak access controls or inadequate validation. Teams evaluating Daybreak-style deployments should require scoped environments, audit logs, expert review, and independent reproduction of model-generated vulnerability findings before remediation or production changes. ## Key Points - 1OpenAI added Blue and Red Daybreak tiers, separating general defensive assistance from advanced vulnerability research and exploit-validation access. - 2OpenAI reported a 95% advanced-request completion rate for GPT-5.6-Cyber, compared with 1.5% for standard GPT-5.6 Sol and 2% through Daybreak Blue. - 3Access is limited to approved users and organizations, with identity checks, monitoring, scoped work, and human oversight among the stated controls. ## Scoring Rationale The program expands controlled access to a purpose-trained cyber model and could affect how enterprise security providers evaluate frontier-model use in defensive workflows. Its exploit-validation capabilities and tiered access design are directly relevant to security engineers, AI governance teams, and practitioners building agent-security tooling. ## Sources Primary source and supporting public references used for this report. Practice interview problems based on real data 1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with. [Try 250 free problems](/problems)