OpenAI is launching GPT-6 Cyber, the latest version of its cybersecurity AI model. The company says it is better at finding vulnerabilities and less likely to go outside its instructions and hack other platforms while doing so.
In OpenAI’s ExploitGym HoneyPot evaluation, GPT-6 Cyber went out of bounds less than 1 percent of the time. GPT-5.6 Cyber did so around 5 percent of the time. The test is modeled on the Hugging Face Incident, in which OpenAI’s AI agents breached the open-source AI community to find answers to a task.
“With the 6 Cyber model, we were able to make huge improvements on this evaluation,” Fouad Matin, head of OpenAI’s cyber preparedness and safety team, said during a press briefing. “It’s over 99 percent of the time doing what you ask.”
Matin said evaluation awareness remains a concern. That is when a model recognizes a test and changes its behavior. “Evaluation awareness is super top of mind for us,” he said. “Often when we do see evaluation awareness, you’ll see it in the model’s chain of thought.” OpenAI is prioritizing production-like evaluations involving real codebases and objectives to lessen that risk, he added.
Read nextGoogle launches Gemini Omni 1.2 Flash with 15-second video clipsOpenAI rates the model Critical for cybersecurity
After a wave of AI cybersecurity incidents, OpenAI briefly paused training its frontier models and now restricts their internet access. Under its Preparedness Framework, the company classifies GPT-6 Cyber as Critical for cybersecurity. The classification means it “can identify and develop functional zero-day exploits of all severity levels in many hardened real-world critical systems without human intervention.” GPT-5.6 Cyber was classified as High.
On ExploitGym, a benchmark measuring whether a model can exploit vulnerabilities, GPT-6 Cyber scored 48.5 percent. That is about 16 points above GPT-5.6 Cyber’s 32.1 percent and above GPT-6 Astra’s 42.4 percent. OpenAI says GPT-6 Cyber beats GPT-6 Astra, its flagship Critical-level model, on most cyber-related tasks.
OpenAI says GPT-6 Cyber helped find 81 previously unknown Firefox vulnerabilities. The model also played a role in hardening the company’s infrastructure for training AI models.
New checks target high-risk activity
OpenAI says GPT-6 Cyber has stronger safeguards “to detect and disrupt misuse” while following user instructions more closely. New checks aim to block “high-risk activity,” including attacks on critical infrastructure such as power plants and water treatment facilities.
GPT-6 Cyber is rolling out to select members of OpenAI’s Daybreak Red program, which serves companies and organizations using its models for “authorized defensive cybersecurity work.” OpenAI plans to expand access to all Daybreak Red customers in the coming weeks.
