News
Public · Published
OpenAI splits Daybreak into Blue and Red tiers, launches GPT-5.6-Cyber
OpenAI split its Daybreak service into two tiers called Blue and Red and launched a new model named GPT‑5.6‑Cyber, which is designed for hacking‑related tasks and claims to answer 95% of advanced cyber prompts.
Published:
Updated:
What happened
OpenAI split its Daybreak service into two tiers called Blue and Red and launched a new model named GPT‑5.6‑Cyber, which is designed for hacking‑related tasks and claims to answer 95% of advanced cyber prompts.
Confirmed
Global impact / market context
The new tiered offering and cyber‑focused model give developers a specialized tool for security testing, potentially increasing demand for AI‑driven penetration testing and raising concerns about misuse by malicious actors.
Analyst inference
AI firms are racing to add niche capabilities as enterprise AI spending grows, while cybersecurity budgets are expanding to counter rising digital threats; OpenAI’s move positions it within both trends.
Analyst inference
What to watch
- Adoption rates of the Red tier by security firms, indicating how quickly the model is being used for professional penetration testing. Analyst inference
- Regulatory responses to AI models that facilitate hacking, which could lead to usage restrictions or reporting requirements. Proposed
- Competitor releases of similar cyber‑oriented AI tools that might erode OpenAI’s first‑mover advantage. Proposed
Affected assets
- GPT — Gold Park