
OpenAI Launches GPT-5.6-Cyber for Cybersecurity Research
The story in brief
OpenAI expanded its Daybreak cybersecurity initiative, offering vetted defenders advanced AI tools like GPT-5.6-Cyber for authorised vulnerability research and security testing. The model outperforms its predecessor, completing 57.3% more requests, and is assessed as reaching the 'High' cybersecurity threshold under OpenAI's framework. Access is restricted to verified individuals and organisations.