tech
OpenAI introduces a new cyber model amid fears of AI cyberattacks
The move comes just days after OpenAI delayed its Astra model over hacking fears.
TL;DR
- OpenAI is introducing GPT-5.6-Cyber, a more permissive AI model for cybersecurity defenders.
- The company is expanding its Daybreak program, which provides access to its cyber models and tools.
- Daybreak will have two tiers: Blue (GPT-5.6 Sol without system guardrails) and Red (GPT-5.6-Cyber for advanced research).
- GPT-5.6-Cyber responded to 95% of advanced cybersecurity requests during testing.
- This release follows the delay of OpenAI's Astra model due to concerns about its hacking abilities.
- Companies like Accenture, IBM, CrowdStrike, Cisco, and Palo Alto Networks can integrate these models into their security products.
- GPT-5.6-Cyber reached the 'High' cyber capability threshold under OpenAI's Preparedness Framework, unlike Astra which reached a higher threshold.