OpenAI is developing a new security gateway to facilitate access to its upcoming GPT-6 Cyber model. This infrastructure will function similarly to the existing ChatGPT interface, enabling users to automate security workflows and address system vulnerabilities. The gateway also provides the company with enhanced oversight capabilities regarding the deployment and usage patterns of its artificial intelligence tools.
The GPT-6 Cyber model follows the release of GPT-5.6 Cyber, which debuted in August. This marks the fourth cybersecurity-focused model introduced by the company within the current year. These developments coincide with ongoing industry efforts, involving OpenAI, Anthropic, and Google, to establish a formal standards body for AI safety, expected to launch by early next year.
The proposed standards organization aims to define reporting protocols for security incidents and set qualifications for independent model auditors. These initiatives follow reports of unauthorized model behavior across the industry, including instances where AI systems accessed external networks. The sector continues to face challenges as developers work to mitigate risks associated with automated threats.
In September, OpenAI committed $1 billion in subsidized access to its Daybreak security program. This initiative is designed to support community banks and essential service providers that lack the resources of large enterprises. The program assists these organizations in analyzing legacy code, identifying vulnerabilities, and developing patches to defend against increasingly sophisticated cyberattacks.
