OpenAI is reportedly preparing GPT-6 Cyber, a new cybersecurity-focused model that could be previewed at DevDay in San Francisco on 29 September. Fortune, citing multiple people familiar with the plans, says the company is also preparing an unnamed product to help customers deploy the model more securely and automatically. Reuters said it could not independently verify the report.
The reported product would support automated security workflows, including vulnerability patching, while giving OpenAI more oversight of how the models are used. OpenAI has not announced the model, the product, its capabilities, pricing or general availability. Read Fortune’s report and Reuters’ follow-up.
GPT-6 Cyber would extend OpenAI’s Daybreak strategy
A limited group of customers is already testing GPT-6 Cyber through Daybreak Red, Fortune reports. OpenAI describes Daybreak Red as its access tier for purpose-trained cybersecurity models used in authorised vulnerability research, exploit validation and security testing. Daybreak Blue is the broader tier for defensive work with general-purpose models.
OpenAI introduced GPT-5.6-Cyber in August and said it was available through Daybreak Red. The company’s earlier security-focused model was limited to select cybersecurity teams, while GPT-6 Astra launched with a separate cybersecurity capability. GPT-6 Cyber would be the fourth cybersecurity-focused model OpenAI has released this year, according to Fortune.
What is confirmed before DevDay?
OpenAI has confirmed the wider Daybreak programme, but not GPT-6 Cyber. Its public programme includes a $1 billion commitment to subsidised Daybreak access, training and technical support for frontline defenders, plus more than 35 partner products and services. OpenAI says thousands of defenders across 2,000 approved organisations and workspaces already use Daybreak.
The reported preview arrives after several incidents involving AI agents escaping test environments and reaching outside websites. Those incidents are part of the backdrop, not proof of what GPT-6 Cyber can do. For now, UAE businesses have no confirmed local pricing, availability or application route for the reported model. DevDay should clarify whether OpenAI is announcing a preview, an alpha expansion or a wider product launch.
Is GPT-6 Cyber officially confirmed?
No. Fortune reported the planned preview, while Reuters said it could not independently verify the report. OpenAI has not publicly announced GPT-6 Cyber.
What is Daybreak Red?
Daybreak Red is OpenAI’s application-only access tier for purpose-trained cybersecurity models used in authorised vulnerability research, exploit validation and security testing.
Can UAE businesses access GPT-6 Cyber?
There is no confirmed UAE pricing, availability date or application route for GPT-6 Cyber. Any local access will depend on OpenAI’s eventual programme and regional terms.
