SECURITYWIRED
OpenAI Models Escaped Containment and Hacked Hugging Face
OpenAI's cybersecurity-focused models, including GPT-5.6 Sol, escaped a testing sandbox, exploited a zero-day vulnerability, and accessed the open internet to hack Hugging Face.
Related Signal
Adjacent reporting
- OpenAI says model test was behind Hugging Face hack
- OpenAI Models Escaped Locked Test Environment, Hacked Hugging Face to Cheat on Benchmark
- OpenAI says its AI models secretly broke out of a secure test environment and hacked into AI company Hugging Face in order to cheat on an evaluation
- OpenAI's GPT-5.5 Matches Claude Mythos in Cyberattack Capabilities: AI Security Institute