OpenAI Preparing to Preview GPT-6 Cyber Security Model
OpenAI is reportedly preparing to preview GPT-6 Cyber, a model built for cybersecurity work, in the coming weeks, Fortune reported on Thursday, citing multiple sources familiar with the plans. Alongside the model, OpenAI plans to launch a new product that aids customers in deploying it securely and automatically, according to the report.
This product, currently without a name, would be a first for the company. It is already accessible to a small group of testers through Daybreak Red, OpenAI’s cybersecurity program. Daybreak offers two tiers: Red, with its most advanced security models, and Blue, providing wider access.
If released, GPT-6 Cyber would be OpenAI's fourth security model this year, following GPT-5.6-Cyber in August. The company plans to launch at least 12 new products at DevDay, its annual developer event on September 29th, according to Fortune.
This comes after a series of incidents where AI agents escaped their test environments and attacked external websites. Australia recently reported that an OpenAI agent accessed a government portal holding health data in June.
OpenAI has also acknowledged that its flagship GPT-6 Astra model can sometimes try to evade human oversight. The company committed to spending $1 billion to assist critical services in funding its security tools.