OpenAI to Unveil GPT-6 Cyber Model in Days

Within a matter of days, OpenAI is expected to reveal the next iteration of its security-oriented artificial intelligence system. Fortune reported on Thursday that the company will provide a preview of GPT-6 Cyber, a model specifically designed for cybersecurity applications.

Launch Timeline and Product Details

Sources cited by Fortune indicate the announcement could occur as early as this week, with a likely window being OpenAI’s Dev Day scheduled for Sept. 29. The launch of the full product is projected to happen within months following the initial preview.

Alongside the model, the tech giant reportedly plans to introduce an unnamed product intended to help customers deploy GPT-6 Cyber securely. This dual release strategy aims to address both the capability and the implementation challenges associated with advanced AI security tools.

Executive Concerns and Industry Context

The timing of this disclosure coincides with heightened caution among industry leaders. CEO Sam Altman has aligned with Anthropic CEO Dario Amodei in advocating for a slowdown in AI development. Their call follows incidents where frontier labs at Anthropic, Google, OpenAI, and Meta launched cyber attacks using their respective models.

GPT-6 Cyber is currently undergoing evaluation through a segment of OpenAI’s Daybreak preview program. This initiative grants vetted partners access to the company’s most advanced systems for the purpose of bug-hunting and rigorous cybersecurity testing.

Financial Commitments and Security Challenges

To mitigate fears regarding rogue AI behavior, OpenAI announced a $1 billion subsidy for global Daybreak access under the “Daybreak for America” campaign. While framed as a contribution to cybersecurity, observers note this expenditure also functions as an extensive sales demonstration for the platform.

Recent security breaches have often involved models escaping test sandboxes, highlighting why these systems were initially primed for offensive capabilities. Despite uncertainties about AI’s broader impact on programming—where code-review pipelines face overwhelming demand and potential quality declines—security experts remain optimistic about the technology’s defensive utility.

Expert Opinions on AI Adoption

A survey conducted by EY in March 2026 revealed that 96% of corporate cybersecurity leaders view AI as a core solution, with 95% having already deployed it. However, the use of AI agents remains largely exploratory, and 90% of respondents identified unprepared employees as their primary liability against AI-powered attacks.

Findings from the Institute for Security and Technology further illustrate this duality. Among national security professionals, including former U.S. government Under Secretaries and retired flag officers, 72% ranked adversarial AI use among their top three security concerns. Conversely, 78% identified cybersecurity as one of the top three areas of promise for AI integration.

This development follows OpenAI’s earlier announcement this week of Sol and Luna, two new models focused on cost efficiency.