We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·May 7·5 min read
Scaling Trusted Access for Cyber with GPT-5.5 and GPT-5.5-Cyber
TL;DR
Security researchers lacked AI tools capable of safely assisting with vulnerability discovery without enabling malicious use. OpenAI expanded its Trusted Access for Cyber program by releasing GPT-5.5 and GPT-5.5-Cyber, specialized models for verified defensive security professionals.
✦ Why It Matters
Defenders can now use specialized AI models to accelerate vulnerability discovery and strengthen critical infrastructure defenses.
Key Takeaways
- Request access to GPT-5.5-Cyber through OpenAI's Trusted Access for Cyber program if your organization conducts defensive security work.
- Use GPT-5.5-Cyber to accelerate vulnerability discovery and analysis workflows within your security research team.
- Understand that access requires verification as a legitimate defender to prevent misuse for malicious purposes.
- Integrate these models into existing vulnerability research pipelines to reduce manual analysis time and improve coverage.
Full Summary
Vulnerability research—the process of identifying and documenting security flaws in software—traditionally requires significant manual effort and expertise. OpenAI addressed this gap by expanding Trusted Access for Cyber, a program that grants verified security defenders early access to advanced AI models.
The expansion introduces GPT-5.5 and GPT-5.5-Cyber, models specifically optimized for cybersecurity applications while maintaining safety constraints. GPT-5.5-Cyber is tailored for defensive security work, enabling researchers to accelerate vulnerability discovery and analysis.
Access is restricted to verified defenders through a vetting process, ensuring the tools support legitimate infrastructure protection rather than malicious activities. This approach balances capability with responsibility, allowing security professionals to leverage cutting-edge AI for defensive purposes while mitigating risks of misuse.