As AI-led attacks multiply, OpenAI launches a new cyber model
OpenAI is putting its new GPT-5.6 Cyber model behind the higher-risk Red tier of Daybreak.
Daybreak now has two tiers, Blue and Red, both limited to approved customers. Blue covers defensive work such as incident response, malware analysis, and patch validation. Red adds purpose-trained cybersecurity models for security testing and vulnerability research, including GPT-5.6 Cyber. OpenAI says the model is available only to trusted customer partners, with Accenture, IBM, CrowdStrike, and Cloudflare reportedly among them. TechCrunch AI's note
Daybreak now has two tiers, Blue and Red, both limited to approved customers. Blue covers defensive work such as incident response, malware analysis, and patch validation. Red adds purpose-trained cybersecurity models for security testing and vulnerability research, including GPT-5.6 Cyber. OpenAI says the model is available only to trusted customer partners, with Accenture, IBM, CrowdStrike, and Cloudflare reportedly among them. TechCrunch AI's note
score 7