OpenAI is preparing to launch GPT-6 Cyber, an artificial intelligence model designed specifically for cybersecurity tasks, according to Fortune magazine, citing sources familiar with the company’s plans.
The new model could be unveiled at DevDay, OpenAI’s annual event scheduled for Tuesday, 29 September, in San Francisco, where the company is expected to present several new products. GPT-6 Cyber itself, however, is expected to become available in the coming weeks, according to Fortune.
According to the company, “the model comes at a time when the ability of AI systems to perform cybersecurity-related tasks has been increasing.” OpenAI has already released several other cybersecurity-oriented models this year, including GPT-5.4 Cyber, GPT-5.5 Cyber, and GPT-5.6 Cyber. GPT-6 Cyber will be, according to Fortune, the fourth cybersecurity-dedicated model the company has launched this year.
The company is also developing a new, as yet unnamed product designed to help clients use and deploy artificial intelligence models in a more secure and automated manner. The tool is expected to enable the creation of automated workflows and the remediation of vulnerabilities, while also giving OpenAI greater capacity to monitor the use of its models.
A limited segment of OpenAI’s clients will already have access to GPT-6 Cyber for alpha testing through the Daybreak Red programme, the company’s initiative dedicated to authorised cybersecurity work. Daybreak also includes a Blue tier, aimed at a broader set of eligible users.
The development of these products comes amid growing scrutiny of the risks associated with the use of AI systems in cyber operations. In September, OpenAI disclosed that GPT-6 Astra, launched earlier in the month, had reached a “Critical” capability level in cybersecurity under its Preparedness Framework.
The company stated that Astra is capable, under certain conditions and with the appropriate tools, of identifying previously unknown security vulnerabilities and developing methods to exploit them. For this reason, OpenAI has reinforced its security measures and restricted access to the model’s most advanced cyber capabilities.
The push also comes as the company seeks to expand the adoption of its security products among organisations and enterprises. According to Fortune, OpenAI has been strengthening its enterprise cybersecurity product offering as part of a strategy that includes new models and tools for authorised use in this domain.
Source: Diário Económico
Original article: https://www.diarioeconomico.co.mz/?p=547896
