OpenAI is getting ready to preview its newest cybersecurity-focused mannequin, GPT-6 Cyber, within the coming weeks, based on a number of sources aware of the plans. It should additionally unveil a brand new product to assist prospects deploy the mannequin in a safer, automated method.
The brand new, yet-to-be-named product will probably be a primary for OpenAI, and extends a paradigm that OpenAI pioneered with its client merchandise, during which ChatGPT serves as a gateway by means of which many customers entry new fashions like GPT-6 Astra. It goals to assist prospects create automated workflows and patch vulnerabilities within the face of more and more refined, AI-powered cyberattacks. It additionally provides OpenAI extra oversight into how its fashions are getting used, to observe for security.
A restricted variety of prospects in OpenAI’s application-only Dawn Purple program have already got entry to GPT-6 Cyber for alpha testing, the supply stated. Dawn is the corporate’s cybersecurity program, which is damaged into two entry tiers—crimson and blue. Clients within the Dawn Purple program have entry to OpenAI’s most superior cybersecurity merchandise. Dawn Blue is its extra generalized entry program, although it’s also application-only.
The discharge of the brand new cyber merchandise comes as OpenAI and different main AI labs have been beneath hearth for a string of worrisome incidents during which “rogue” brokers escaped their sandboxes and hacked exterior web sites together with the Hugging Face web site and an Australian authorities web site. However whereas the incidents have set off alarms about AI’s safety dangers, OpenAI’s new cybersecurity instruments are a part of an ongoing effort by the corporate to develop AI safety merchandise.
GPT-6 Cyber is the fourth cybersecurity-focused mannequin OpenAI has launched this yr. Its first was GPT-5.4 Cyber in April 2026, adopted by GPT-5.5 Cyber in June, and GPT-5.6 Cyber in August. OpenAI has been specializing in enterprise gross sales of cybersecurity merchandise, an effort now led by its Chief Income Officer Dali Rajic, who joined the corporate in August.
OpenAI stated earlier this month it’ll make investments $1 billion to subsidize the usage of its cybersecurity merchandise for essential companies.
Subsequent week, OpenAI additionally plans to ship a dozen or extra different merchandise at its annual DevDay occasion, set for Sept. 29, based on one other supply aware of this system. Most of these releases aren’t associated to cybersecurity however quite different areas of the enterprise, each enterprise- and consumer-focused.
OpenAI has put a freeze on any main launches prior to now 2 weeks, save for its affordability-focused GPT-6 Sol and Luna fashions, so it might ship all of them at DevDay, the supply stated. Earlier this month, CEO Sam Altman teased the buffet of recent releases at DevDay in a tweet with a number of ship emojis.
UPDATE: The story has been up to date to mirror that the launch might be inside weeks, not days.
