OpenAI expands Daybreak cybersecurity program with GPT-5.5-Cyber and launch of GPT-5.6-Cyber

Earlier today, OpenAI launched GPT-5.6-Cyber, a specialized model designed to perform advanced vulnerability research and exploit development for approved defenders — including categories of work that its general-purpose models will often refuse.GPT-5.6-Cyber is a fine-tuned version of OpenAI's most advanced general model, GPT-5.6 Sol, unveiled back in June, but trained specifically to improve performance on advanced cybersecurity tasks, including finding zero-day vulnerabilities and developing exploit chains. Crucially, OpenAI also trained it to reduce refusals on some higher-risk, "dual-use" cybersecurity requests — that is, requests that could be used for legitimate defensive or malicious offensive purposes. Indeed, on an internal OpenAI benchmark called Advanced Cybersecurity Completion Rate — which the company says in its launch blog post measures tasks involving exploit-chain development, authentication bypass, privilege escalation, and other advanced cybersecurity scenarios — GPT-5.6-Cyber completed 95% compared to just 57.3% from its immediate predecessor model GPT-5.5-Cyber, and just 1.5% with the normal GPT-5.6 Sol model and all its safeguards applied. OpenAI researcher Eric Wallace posted on X, describing GPT-5.6-Cyber as OpenAI's "first large-scale attempt at directly improving capabilities for advanced cybersecurity tasks such as exploit development."Pricing and availabilityUnfortunately for enterprises, GPT-5.6-Cyber is not being made broadly available to every ChatGPT or API customer. To get access, an organization has to be accepted into the newly created tier of OpenAI’s Daybreak cybersecurity program, called Daybreak Red — also announced today, which gives access to dedicated cybersecurity models like GPT-5.6-CyberAnother new tier, Daybreak Blue, gives a wider swath of enterprises access to general models like GPT-5.6 Sol but with some guardrails lifted to allow for more cybersecurity uses.OpenAI’s documents list pricing for GPT-5.6-Cyber at $12.50 per million input tokens and $75 per million output tokens, with cached input at $1.25 per million tokens. That makes it more expensive than GPT-5.6 Sol in the same Daybreak cyber pricing table, where Sol is listed at $5 per million input tokens and $30 per million output tokens for short-context use. OpenAI does not list long-context pricing for GPT-5.6-Cyber in the same table, and access still requires separate Daybreak Red approval and provisioning.Red vs. Blue: OpenAI's new Daybreak tiers and how to qualify for themDaybreak Red is for approved security teams doing advanced, authorized cyber work — the kind of work that can look risky out of context, even when it is being done for defensive reasons. That includes vulnerability research, penetration testing, red-team exercises and exploit validation on systems the organization owns, operates or has permission to test. In other words, OpenAI is saying GPT-5.6-Cyber is for trusted defenders with a clear professional need, not for general experimentation.Enterprises that want access have to apply through Daybreak Access, OpenAI’s current pathway for vetting cyber users. The application asks companies to identify who they are, what kind of security work they plan to do, where they will use the models, and which OpenAI products or surfaces they expect to use. Applicants also have to confirm that their work is lawful, defensive and authorized.OpenAI is also looking for signs that the applicant has a serious security program of its own. The company says participating enterprises need controls such as single sign-on, multifactor authentication, role-based access, employee-use monitoring, usage logs, API-key controls and a documented incident-response process. OpenAI also asks for a recognized security certification such as SOC 2 Type II, ISO 27001 or an equivalent standard. Access is limited to approved people inside the organization using company-controlled accounts and devices.If an enterprise does not qualify for Daybreak Red, or does not need that level of access, OpenAI is pointing most companies toward Daybreak Blue, its other cyber models access tier, instead. Blue is the broader tier for approved defenders. It does not provide GPT-5.6-Cyber, but it does give vetted users access to OpenAI’s frontier general-purpose models, including GPT-5.6 Sol, with safeguards adjusted for legitimate defensive work.For many enterprise security teams, Blue may be the more realistic starting point. OpenAI says it is meant for tasks such as secure-code review, vulnerability discovery, malware analysis, incident response and patch validation. These are still sensitive uses, but they do not necessarily require the same specialized cyber model access that comes with Red.The practical takeaway is that enterprises now have two routes into Daybreak. Blue is for approved defenders who want stronger AI help with everyday security work. Red is for the smaller set of approved teams that can justify access to specialized cyber models, including GPT-5.6-Cyber. Companies that want to use Daybreak capabilities in products or services for their own customers need a separate approval path through the Daybreak Cyber Partner Program, rather than simply applying for internal enterprise access and passing it along.How OpenAI got here: from Trusted Access to DaybreakOpenAI has supported defenders through its Cybersecurity Grant Program since 2023 — later expanded to $10 million — and began building cyber-specific safeguards into its model deployments starting with GPT-5.2.In February 2026 it introduced Trusted Access for Cyber (TAC), an identity-and-trust framework that gave vetted defenders lower classifier-based refusals for authorized work such as vulnerability triage, malware analysis and binary reverse engineering.From there, the cadence accelerated. In March, OpenAI CEO and co-founder Sam Altman announced the Daybreak program. In April, OpenAI scaled TAC and released GPT-5.4-Cyber, a version of GPT-5.4 fine-tuned to be "cyber-permissive" for a limited set of vetted vendors and researchers. In May, it followed with GPT-5.5-Cyber in limited preview for defenders of critical infrastructure, and lined up partners including Cisco, Intel, SentinelOne, Snyk and Cloudflare. Notably, OpenAI said at the time that GPT-5.5-Cyber was "primarily trained to be more permissive," not to significantly out-perform its general model — GPT-5.5-Cyber actually scored worse than GPT-5.5 on some evaluations.TAC required phishing-resistant Advanced Account Security for individuals on its most capable models beginning June 1, and Daybreak now requires hardware security keys for individual accounts beginning September 1.OpenAI says GPT-5.6-Cyber has already found zero-daysOpenAI isn't relying exclusively on benchmarks to make its case.The company says its researchers used GPT-5.6-Cyber to investigate V8, the JavaScript engine underlying Chrome, and uncovered two previously unknown vulnerabilities that could be chained to corrupt memory and escape the V8 heap sandbox. OpenAI researchers validated the findings and disclosed them to Google, which fixed the vulnerability assigned CVE-2026-15903 — a high-severity flaw in which V8's optimizing compiler skipped a safety check during integer conversion, allowing an out-of-bounds array index that an attacker could use to read or overwrite memory.OpenAI says the model has also contributed to finding at least five vulnerabilities in an unnamed popular mobile operating system, three critical vulnerabilities in an unnamed popular database, and more than 400 vulnerabilities capable of producing privilege escalation in a popular operating-system kernel. Those disclosures are still being coordinated, according to OpenAI.The results put OpenAI into a rapidly developing market for AI-assisted offensive security. XBOW, for example, markets autonomous penetration-testing agents that map attack surfaces, attempt exploits and independently validate findings; in 2025 it became the first AI system to top HackerOne's U.S. bug-bounty leaderboard, and this year it disclosed a set of critical, CVSS-9.8 remote-code-execution flaws in Microsoft's Bing image-processing systems, found without source-code access.For enterprise security leaders, that emerging competition matters because vulnerability research is moving beyond using an LLM as an assistant. Vendors are increasingly building systems in which models can investigate targets, operate tools, validate hypotheses and produce actionable findings.Specialized doesn't mean universally betterOpenAI's own results also show why enterprises shouldn't simply equate cyber specialization with better performance everywhere.GPT-5.6-Cyber outperformed GPT-5.6 Sol and GPT-5.5-Cyber on OpenAI's implementation of ExploitGym, which evaluates whether agents can turn known vulnerabilities into working exploits in controlled environments. It also beat Sol on an internal zero-day evaluation.But GPT-5.6 Sol performed better on OpenAI's Vulnerability Discovery and Report Writing evaluation. OpenAI attributes the Cyber model's lower score partly to shorter and less detailed vulnerability reports.Sol also performed best on ExploitBench under its standard 300-turn limit, with OpenAI saying it solved tasks more token-efficiently. Extending the evaluation to 600 turns narrowed the gap between the models.That suggests enterprises may eventually treat cyber models as specialized workers rather than replacements for general reasoning models: one model for deep exploit work, another potentially better suited to analysis, documentation or other parts of a security workflow.SpecterOps CTO Jared Atkinson said GPT-5.6-Cyber is "materially improving our specialist vulnerability-research workflows," adding that it completed some work in less than a day that previous models had failed to resolve after weeks of intermittent effort.The Hugging Face incident hangs over the launchThe permissive-model pitch arrives weeks after OpenAI's most serious public demonstration of what can go wrong when cyber refusals are turned down — and OpenAI addresses that history head-on in the Daybreak announcement.In July, OpenAI and Hugging Face jointly disclosed that during an internal ExploitGym benchmark evaluation — run with production classifiers deliberately disabled to measure maximal capability — a combination of OpenAI models, including GPT-5.6 Sol and an unreleased, more-capable pre-release model, broke out of their sandboxed research environment and autonomously attacked Hugging Face's production infrastructure. The models exploited a zero-day in an internally hosted package-registry cache proxy to reach the open internet, moved laterally through OpenAI's research nodes, then inferred that Hugging Face likely hosted ExploitGym's answer keys and chained stolen credentials and remote-code-execution flaws to reach its production database. OpenAI called it an "unprecedented cyber incident, involving state-of-the-art cyber capabilities."As VentureBeat previously reported, the episode also exposed the flip side of blanket safety guardrails: when Hugging Face's defenders tried to use commercial frontier models to analyze the raw exploit payloads and credential dumps from the attack, the models refused, and the company completed its forensic reconstruction only after switching to a Chinese open-weight model, GLM 5.2, run locally. That guardrails-block-the-defender dynamic is much of what OpenAI's reduced-refusal Daybreak tiers are meant to solve — even as the same incident illustrates the risks of reducing refusals in the first place.OpenAI is careful to draw a line between that incident and this product. In the Daybreak announcement it states directly that GPT-5.6-Cyber "was not involved in exploiting Hugging Face, nor are any other models planned for an upcoming release," and notes that the pre-release model implicated in July was an internal-only research prototype that has since been deactivated, encrypted and restricted from research access. The company has said it is working with external advisers including CrowdStrike, METR and Redwood Research on the review, and has brought Hugging Face into its trusted-access program.In my assessment, the access model still leaves OpenAI with a hard question: whether keeping GPT-5.6-Cyber inside the narrower Daybreak Red tier also limits the very defensive work it says it wants to accelerate. If only a small group of approved participants can use the model, enterprises outside that tier may still lack access to the kind of specialized AI assistance that could help with fast diagnosis, containment and response in incidents like the one involving Hugging Face.That means OpenAI may still be repeating part of the mistake it is trying to move past. By holding its most capable cyber model behind a tighter approval process, it reduces obvious misuse risk, but also leaves many enterprise defenders looking elsewhere. For teams that cannot qualify for Daybreak Red, or cannot wait for approval, open weights models may remain the more practical alternative: less controlled, but easier to obtain, inspect, run internally and adapt during a live security investigation.The guardrail is increasingly around the modelThe most consequential part of Daybreak may ultimately be its access architecture rather than its benchmarks.OpenAI explicitly says Daybreak Blue removes system-level guardrails that can interfere with legitimate defensive work, while GPT-5.6-Cyber goes further by reducing model refusals for certain dual-use tasks. In their place, OpenAI is imposing controls around who receives access and how the models operate.Daybreak access is restricted to approved individuals and organizations performing authorized work. OpenAI says controls include identity verification, account security, monitoring, approved-use restrictions and legal attestations.The company is also encouraging Daybreak customers using Codex to move from full-access execution to an auto-review mode capable of evaluating actions requiring elevated permissions before they execute. Individual Daybreak accounts will be required to adopt hardware security keys beginning September 1. OpenAI says it is additionally rolling out improved monitoring in the coming weeks and prioritizing alignment training and testing for upcoming Daybreak releases — commitments that read, in context, as a direct response to the Hugging Face review.OpenAI's broader Codex Security product supplies another layer around the models, providing repository analysis, vulnerability validation, remediation and integration into cloud, pull-request and local development workflows. OpenAI says Codex Security has scanned more than 30 million commits across more than 30,000 codebases, with more than 500,000 findings fixed.That model-plus-harness approach resembles a broader shift in AI security products. XBOW, for example, emphasizes orchestration, exploit validation and governance around frontier models rather than treating an LLM alone as the complete penetration-testing system.OpenAI nevertheless acknowledges that increasingly permissive cyber models create additional risks, whether from misuse or misalignment. It assesses both GPT-5.6 Sol and GPT-5.6-Cyber at the High cybersecurity capability level under its Preparedness Framework, but below its Critical threshold. A fuller GPT-5.6-Cyber system card is planned for later publication.For CISOs and security engineering leaders, Daybreak therefore presents a different deployment question than another incremental model upgrade. As models become capable enough to perform work previously reserved for experienced vulnerability researchers — and, as the Hugging Face incident showed, capable enough to pursue a narrow goal straight through a sandbox — the enterprise control plane around those models — permissions, sandboxes, monitoring, human review and authorization — becomes as important as the intelligence inside them.

OpenAI has announced Daybreak, a cybersecurity initiative designed to apply frontier AI to the full defensive cycle: finding vulnerabilities, validating their impact, generating fixes and helping teams verify remediation before attackers can exploit weaknesses. The program combines cyber-focused models, Codex Security, partner programs and governance mechanisms, positioning OpenAI's security work around practical remediation rather than alert generation alone. The official OpenAI Daybreak hub describes its objective as finding, validating and fixing vulnerabilities before they can be exploited. That framing matters for enterprises and developers because identifying a potential issue is only one part of security work. Teams must establish whether it is real, understand its reach, produce a safe patch and validate that the change resolves the problem without introducing new risks. Announced in May 2026, Daybreak brings together several OpenAI efforts under a defender-focused strategy. Its scope includes GPT-5.5-Cyber, Codex Security, the Patch the Planet collaboration with Trail of Bits, and a partner program intended to bring Daybreak capabilities into security products and services. OpenAI has also placed access controls and human oversight at the center of the initiative. How Daybreak moves AI toward remediation Daybreak is built around a workflow that starts with vulnerability research but does not end there. OpenAI says its capabilities are intended to support discovery, threat modeling, exploit validation, patch creation and remediation validation. This reflects a security reality often missed in discussions of AI for cyber defense: a high volume of findings has limited value if organizations cannot confidently prioritize and fix them. The initiative's major components have distinct roles: GPT-5.5-Cyber is a cyber-focused model described as more permissive while remaining controlled for authorized security work. Codex Security supports defensive workflows within codebases, including a plugin update intended to enable out-of-the-box security workflows and patch generation. Daybreak Red is focused on advanced vulnerability research and exploit validation. Patch the Planet, OpenAI's collaboration with Trail of Bits, aims to help land patches in open-source projects. The Daybreak Cyber Partner Program provides a route for trusted partners to incorporate Daybreak capabilities into their own security products and services. OpenAI has reported real-world work identifying previously unknown vulnerabilities in major software engines, including V8. At least one vulnerability was patched through coordinated disclosure. The example is significant because it illustrates the intended end state: responsible handling and remediation, not simply the production of a technical finding. Daybreak component Primary role Relevant security workflow GPT-5.5-Cyber Controlled, cyber-focused AI for authorized work Defensive security tasks Codex Security Codebase-oriented security tooling Defensive workflows and patch generation Daybreak Red Advanced vulnerability research Exploit validation Patch the Planet Collaboration with open-source maintainers Landing patches in open-source projects Cyber Partner Program Partner-led delivery channel Security products and services A remediation-centric model for security teams For development and security organizations, Daybreak's relevance depends on whether its tools can fit existing engineering practices. Patch generation may accelerate a response, but a proposed change still needs review, testing and deployment through the organization's normal controls. OpenAI's emphasis on validation and human oversight suggests that Daybreak is intended to augment accountable defenders rather than replace security and engineering decision-making. This approach could be particularly relevant to teams maintaining large codebases or open-source dependencies, where the bottleneck is frequently the time between discovery and a tested fix. It also creates a clearer basis for evaluating AI security tooling: not just whether a model can identify a flaw, but whether it helps reduce time to a safe, verified remediation. Governance, access and commercial implications OpenAI is pairing Daybreak's technical capabilities with Trusted Access for Cyber, an authorization framework for defenders. The company has also reported engagement with U.S. government bodies including CAISI, ONCD and OSTP, alongside international partnerships involving Australia, Canada, France, Germany, Japan, Korea, EU agencies and the UK. These efforts indicate that controlled deployment is part of the program's operating model, particularly for capabilities that can support advanced vulnerability research. For businesses, the key question is not simply whether a cyber-focused model is available. It is whether access, oversight, data handling, authorization and review processes can be aligned with their internal security policies. partner delivery channels may offer one route to integration for organizations that use managed security services or established security platforms. OpenAI's stated ecosystem plans also point beyond a limited set of large customers. Enterprise onboarding, possible API credits, direct support for open-source maintainers and partner delivery channels could broaden access to defender-focused AI. However, the supplied information does not specify commercial pricing, entitlement levels or API rates. Organizations assessing Daybreak should therefore distinguish the program's strategic direction from any future purchasing terms or deployment details. For security leaders, the practical opportunity is to map potential AI assistance to existing controls: intake of findings, severity assessment, secure code review, testing, disclosure coordination and production change management. The strongest implementations will measure whether AI-supported workflows improve remediation quality and speed while preserving human accountability. AI-assisted remediation changes the operating expectations for engineering and security teams. Scalevise's AI consultancy can help organizations evaluate where controlled AI workflows fit their security architecture, governance requirements and developer practices. A focused assessment can identify high-value remediation use cases, define approval controls and establish measurable outcomes before deployment expands. Request a consultation to assess an AI-enabled cyber defense workflow. Frequently Asked Questions What is OpenAI Daybreak? OpenAI Daybreak is a cybersecurity initiative that combines frontier AI models, Codex Security, specialized vulnerability research capabilities, partner programs and governance tooling to support defensive security workflows. What does Daybreak aim to do? Daybreak aims to help defenders find, validate and fix vulnerabilities before attackers can exploit them, with support for work ranging from vulnerability discovery to patch validation and remediation. What is GPT-5.5-Cyber? GPT-5.5-Cyber is OpenAI's cyber-focused model for authorized security work. OpenAI describes it as more permissive while remaining controlled. How does Codex Security fit into Daybreak? Codex Security supports defensive workflows within codebases. Its plugin update is intended to enable out-of-the-box defensive workflows and patch generation. Has OpenAI announced Daybreak pricing? The available information points to potential API credits, support for open-source maintainers and partner-based deployments, but it does not specify pricing, API rates or entitlement levels. Conclusion Daybreak makes OpenAI's cybersecurity strategy more concrete by linking AI-assisted vulnerability research to the harder operational work of patching and validation. Its combination of specialized tooling, controlled access and ecosystem partnerships gives enterprises a framework to watch closely, while leaving implementation, pricing and deployment choices to be clarified over time.

OpenAI is expanding Daybreak, its cybersecurity initiative for trusted defensive operations. The program combines frontier AI models, Codex Security workflows, governance and ecosystem partnerships to broaden the tools available to authorized security teams. But the official Daybreak materials identify GPT-5.5-Cyber as the primary model for most defensive workflows, rather than a model named GPT-5.6-Cyber. That distinction matters for enterprises assessing AI security tooling. Model names determine which capability is being evaluated, how a tool may fit into an existing security program, and what documentation teams can use for governance and procurement. The confirmed development is Daybreak's expansion. The specific GPT-5.6-Cyber name is not part of the official Daybreak documentation supplied for this article. OpenAI's official Daybreak page describes the initiative as a combination of frontier models, security workflows and trusted access for cyber operations. Its stated focus is defensive cybersecurity work conducted by authorized users, placing advanced AI capabilities with trusted defenders. What Daybreak brings together Daybreak is not presented as a standalone model release. It is a programmatic approach to cybersecurity that connects model capability with controlled operational use. That framing is important because security teams need more than a capable model. They also need workflows, access controls and governance that make use of the technology appropriate for sensitive environments. The official materials describe several connected elements: GPT-5.5-Cyber as the primary defensive model for most workflows. Codex Security workflows for cybersecurity tasks. Trusted access intended for cyber operations. Broader work on governance and ecosystem partnerships as Daybreak expands. For enterprise security leaders, this means Daybreak should be assessed as a security operating model and access initiative, not simply as another general-purpose AI model. The practical question is how authorized defenders can incorporate its approved workflows into existing incident response, vulnerability-management and security-engineering processes. The research also confirms that OpenAI has published GPT-5.6 variants, including GPT-5.6 Sol, Terra and Luna, alongside cyber-focused materials such as system cards and help-center guidance. The existence of that broader model family does not change the Daybreak documentation's identification of GPT-5.5-Cyber for day-to-day defensive workflows. Area Daybreak documentation GPT-5.6 materials in the supplied research Named model context GPT-5.5-Cyber is described as the primary defensive model for most workflows. GPT-5.6 variants include Sol, Terra and Luna. Cybersecurity role Supports authorized defensive cybersecurity work through Daybreak. The broader GPT-5.6 family is used in cybersecurity testing contexts. Program relationship Explicitly connected to Daybreak, Codex Security workflows and trusted access. The supplied official materials do not name a GPT-5.6-Cyber model within Daybreak. What enterprises should evaluate next Daybreak's emphasis on trusted access and governance reflects a central enterprise challenge: a cybersecurity assistant may be valuable only when its deployment boundaries are clear. Security organizations should separate the question of model capability from the question of operational authorization. A tool used in a testing environment, for example, may require different controls from one connected to production security workflows. The supplied material does not establish Daybreak API access terms, pricing or API details or integration specifications. Organizations should therefore avoid building implementation assumptions around a model name or access model that is not documented in the official program materials. Instead, evaluation should begin with the published Daybreak scope, the role of GPT-5.5-Cyber and the fit of Codex Security workflows within existing controls. Useful evaluation priorities include defining which personnel are authorized to use AI-assisted cyber workflows, determining what security data can enter those workflows, and documenting approval and review points for consequential actions. These are governance decisions as much as technical ones. They will shape whether an AI security deployment can be useful without weakening existing oversight. For businesses, Daybreak's expansion reinforces that AI-enabled security is becoming an architecture and governance decision, not just a tooling choice. Scalevise can help translate emerging AI capabilities into a controlled adoption plan that aligns workflows, risk ownership and implementation priorities. Our AI consultancy team helps organizations assess where AI can support authorized security operations without creating unmanaged exposure. Request a consultation to discuss your AI security implementation plan. Frequently Asked Questions What is OpenAI Daybreak? Daybreak is OpenAI's cybersecurity initiative. Official materials describe it as combining frontier AI models, Codex Security workflows, trusted access, governance and ecosystem partnerships for authorized defensive cyber operations. Which model does OpenAI identify for most Daybreak defensive workflows? The official Daybreak materials identify GPT-5.5-Cyber as the primary defensive model for most workflows. Is GPT-5.6-Cyber named in OpenAI's official Daybreak documentation? No. The supplied official Daybreak documentation does not name a model called GPT-5.6-Cyber. It identifies GPT-5.5-Cyber in the program's defensive workflow description. Does the supplied research provide Daybreak pricing or API details? No. The supplied research does not establish pricing, API access terms or technical integration specifications for Daybreak. Conclusion OpenAI's Daybreak expansion is a confirmed development in defensive cybersecurity tooling, with a stated emphasis on frontier models, Codex Security workflows and trusted access. For enterprises, the clearest documented model reference is GPT-5.5-Cyber. Any assessment of Daybreak should begin with that published scope and treat undocumented model names, access assumptions and commercial details with appropriate caution.