OpenAI is gearing up to preview GPT-6 Cyber, a cybersecurity-focused artificial intelligence model, within weeks, according to Fortune.
The ChatGPT maker is also working on a companion product to help customers deploy the model safely and automate their security work.
The as-yet-unnamed tool would let companies build automated workflows and patch software vulnerabilities before attackers find them.
It would also give OpenAI closer oversight of how its models are used, handy when the same technology can pick locks as well as fit them.
AI-powered cyberattacks are growing more sophisticated, and defenders are under pressure to fight fire with fire.
A small group of customers is already testing GPT-6 Cyber through Daybreak Red, OpenAI's application-only cybersecurity programme.
In security jargon, red teams play the attacker, probing systems for weaknesses.
Daybreak also has a broader Blue tier, named after the defenders who keep the lights on.
The new model would be OpenAI's fourth cyber-focused release this year.
GPT-5.4 Cyber arrived in April, followed by GPT-5.5 Cyber in June and GPT-5.6 Cyber in August.
That is roughly one every two months, a pace suggesting OpenAI sees security as a serious commercial opportunity rather than a side project.
The company plans to unveil the new products at its annual DevDay developer conference on 29 September.
It could launch a dozen or more products at the event, spanning consumer and business offerings.
For corporate security chiefs, the pitch is simple: if the bad guys are using AI, you had better be using it too.
Whether handing more of the job to the machines makes networks safer, or simply raises the stakes, is a question the industry has yet to answer.
