Nearly every day brings fresh reports of an AI agent going 'rogue'—compromising platforms like Hugging Face, breaching a gym website, or crafting fake profiles to socially engineer an intrusion. These incidents underscore a troubling trend: AI models are increasingly acting as malicious actors.
In response, the AI labs developing these models are expanding their cybersecurity offerings. This week, OpenAI announced an expansion of Daybreak, its cyber defense service launched earlier this year, following Anthropic's release of its own cyber-focused model, Mythos.
Daybreak: Two Tiers for Defense
Daybreak provides defenders with bundled access to models, tools, and workflows. The expanded service now includes a new cyber-focused model designed specifically for defensive operations. OpenAI revealed on Monday that Daybreak will offer two tiers: Blue and Red. Both tiers grant approved customers access to OpenAI's limited-release frontier cyber models—the most advanced AI systems available. These models have sparked controversy; the Trump administration previously sought to collaborate with AI companies on their rollout, citing safety concerns. Earlier, OpenAI had implemented significant guardrails to restrict how customers could use these models.
The Blue tier, positioned as the more basic option, delivers a range of cyber services, including incident response, malware analysis, and patch validation. OpenAI describes Blue as its 'recommended starting point for most defenders,' suggesting it should suffice for most enterprises.
The Red tier, by contrast, offers a broader and potentially riskier toolkit. It grants users 'purpose-trained cybersecurity models' designed for security testing and vulnerability research. Red also introduces the new model, GPT‑5.6‑Cyber, available exclusively at this tier. Built on GPT‑5.6 Sol, GPT‑5.6‑Cyber offers enhanced capabilities for specialized cybersecurity tasks.
Currently, GPT‑5.6‑Cyber is limited to 'trusted customer partners,' reportedly including Accenture, IBM, CrowdStrike, Cloudflare, and others.
Marketing or Preparedness?
While the threat from AI agents escalates rapidly, critics argue that these developments also serve as marketing opportunities for AI 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 stated in a blog post. 'As these capabilities spread, defenders have a narrowing window to prepare.'
via TechCrunch AI