OpenAI has announced an expansion of Daybreak, its cyber defense service. The company originally launched the offering earlier this year. This update follows shortly after Anthropic released its own cyber-focused model called Mythos. Daybreak bundles access to models, tools, and workflows tailored for defenders. The latest expansion introduces access to a brand new cyber-focused model designed specifically for defensive security work.
Bad actors are increasingly using artificial intelligence in malicious ways. Recent incidents include autonomous agents compromising Hugging Face, hacking a gym website, and creating fake profiles to execute social engineering intrusions. Because these autonomous models are behaving like cyber criminals, the major artificial intelligence labs that build them are rapidly expanding their protective security offerings. OpenAI and other creators are rushing to deliver tools that can counteract these emerging threats.
OpenAI stated on Monday that Daybreak will now feature two distinct tiers named Blue and Red. Both tiers provide approved customers with access to the company's limited-access frontier cyber models. Frontier models represent the most advanced technology currently available. These powerful systems have previously drawn significant controversy. The Trump administration sought to collaborate with artificial intelligence companies regarding the rollout of such models due to safety concerns. OpenAI has historically placed heavy guardrails on these technologies to restrict unauthorized use.
The Blue tier serves as the more basic option for enterprise users. It provides various standard security services such as incident response, malware analysis, and patch validation. OpenAI officially designates Blue as the recommended starting point for most defenders. The company implies this tier is sufficient for the everyday security needs of typical enterprises. Most organizations will likely find these baseline defensive workflows adequate for their normal operations.
The Red tier delivers a much broader and potentially more dangerous toolkit. OpenAI grants users in this tier access to purpose-trained cybersecurity models designed for security testing and vulnerability research. The Red tier includes the newly unveiled GPT-5.6-Cyber model. This new model is built upon the GPT-5.6 Sol architecture and provides enhanced capabilities for specialized cybersecurity tasks. At this moment, OpenAI is limiting access to GPT-5.6-Cyber exclusively to trusted customer partners.
Several major firms currently have access to this restricted software. Trusted partners include Accenture, IBM, Crowdstrike, and Cloudflare. While threats from artificial intelligence agents are rising quickly, industry critics note that these security products also act as major marketing opportunities for labs. OpenAI highlighted the changing landscape in a public blog post. The company stated that threat actors will increasingly use artificial intelligence to conduct attacks at unprecedented speed and scale in fully autonomous ways. OpenAI warned that defenders have a narrowing window to prepare as these capabilities spread across the digital ecosystem.
Enterprises continue to show strong interest in purchasing protection directly from artificial intelligence labs. Companies prefer this approach because these labs understand the associated security risks first-hand. Security teams rely on these specialized services to maintain parity with rapidly advancing adversarial tactics. Approved customers will continue evaluating the newly expanded Blue and Red tiers of Daybreak as threat actors adopt more automated methods.
