OpenAI has launched a new cyber model, GPT‑5.6‑Cyber, to counter attacks by autonomous AI agents.
The company announced the expansion of Daybreak, its cyber defense service, on Monday. This move follows similar releases from competitors like Anthropic, whose Mythos model also focuses on security.
Daybreak now splits into two tiers: Blue and Red. Both levels grant access to OpenAI’s limited-access frontier cyber models. These advanced systems have drawn controversy, prompting the Trump administration to seek collaboration with AI firms on their rollout due to safety concerns. Previously, OpenAI applied strict guardrails to restrict how customers could use these tools.
Two levels of access
Blue serves as the recommended starting point for most defenders. It covers incident response, malware analysis, and patch validation. OpenAI suggests this tier provides sufficient coverage for the majority of enterprises.
Red offers a broader and potentially more dangerous toolkit. Users gain access to purpose-trained cybersecurity models designed for security testing and vulnerability research.
The new GPT‑5.6‑Cyber model is exclusive to the Red tier. Built on GPT‑5.6 Sol, it delivers enhanced capabilities for specific cybersecurity tasks.
Currently, GPT‑5.6‑Cyber is available only to trusted customer partners. Reported clients include Accenture, IBM, Crowdstrike, and Cloudflare.
Critics note that the rise in AI-led threats often doubles as a marketing opportunity for these labs. OpenAI is certainly positioning its upgraded Daybreak in that light.
“The cybersecurity world is rapidly changing—threat actors will increasingly use AI to conduct cyberattacks at unprecedented speed and scale, including in fully autonomous ways,” the company said in a blog post. “As these capabilities spread, defenders have a narrowing window to prepare.”
Enterprises remain keen to purchase protection from the same labs that created the models, trusting them to understand security risks from first-hand experience.
