OpenAI Expands Daybreak as AI-Led Cyberattacks Multiply

OpenAI is expanding Daybreak, its AI cybersecurity defense program, and rolling out a new cyber-trained AI model with it. The new model, GPT-5.6-Cyber, is described as a cyber-focused model designed for defensive work and specialized cybersecurity tasks. The move comes as AI agents are increasingly being discussed in connection with rogue or malicious behavior.

OpenAI said Daybreak will now include two tiers: Blue and Red. Both tiers give approved customers access to OpenAI’s limited-access frontier cyber models. Blue includes services such as incident response, malware analysis, and patch validation, and OpenAI calls it the recommended starting point for most defenders. Red offers a broader toolkit for security testing and vulnerability research, including purpose-trained cybersecurity models.

GPT-5.6-Cyber is available only through the Red tier. TechCrunch reports that the model is currently being made available to “trusted customer partners,” including reportedly Accenture, IBM, CrowdStrike, Cloudflare, and others. The restricted rollout reflects the sensitivity around frontier cyber models and the guardrails OpenAI has previously placed on their use.

The article frames OpenAI’s move against a backdrop of rising AI-led security incidents, including examples involving Hugging Face, a gym website, and social engineering activity. OpenAI said threat actors will increasingly use AI to conduct cyberattacks at “unprecedented speed and scale,” including fully autonomous attacks. At the same time, TechCrunch notes that critics see these threats as marketing opportunities for AI labs, even as enterprises remain interested in buying protection from the companies building the models.