I am a Member of Technical Staff at OpenAI. I currently work on the alignment and core capabilities of cybersecurity agents, and have recently been shipping models in OpenAI’s Cyber series, including GPT-5.5 Cyber, GPT-5.6 Cyber and GPT-6 Cyber. I also did some work around prompt injection and automated red-teaming at OpenAI. Earlier, I obtained my Ph.D. at Princeton University, where I was fortunate to be advised by Prof. Prateek Mittal and Prof. Peter Henderson, who are incredibly wise and kind.
Recent Work
- GPT-6 Cyber: Advancing the frontier of aligned cybersecurity agents
- The Next Challenge for Agentic Cybersecurity: A Realistic, Contamination-Free Reverse Engineering Benchmark
- ExploitGym: Can AI Agents Turn Security Vulnerabilities into Real Attacks?
- GPT-Red: Automated Red Teaming via Self-Play at Scale
- Continuously hardening ChatGPT Atlas against prompt injection attacks
Xiangyu Qi, Ashwinee Panda, Kaifeng Lyu, Xiao Ma, Subhrajit Roy, Ahmad Beirami, Prateek Mittal, Peter Henderson
Paper Code
Yannic Kilcher’s Paper Explainer Center for Democracy and Technology letter to NIST
Xiangyu Qi$^* $, Yi Zeng$^* $, Tinghao Xie$^* $, Pin-Yu Chen, Ruoxi Jia, Prateek Mittal, Peter Henderson
Paper Policy Brief Code Website
The New York Times PCMag The Register VentureBeat
GPT-4V(ision) system card cited this paper to underscore the emerging threat vector of multimodal jailbreaking.
Xiangyu Qi$^* $, Kaixuan Huang$^* $, Ashwinee Panda, Peter Henderson, Mengdi Wang, Prateek Mittal
Paper Code