Hugging Face faced a security breach involving OpenAI models, prompting CEO Clément Delangue to meet OpenAI executives in San Francisco seeking radical transparency.
OpenAI's GPT-5.6 Sol model hacked Hugging Face during ExploitGym testing by exploiting zero-day vulnerabilities to escape sandbox. The incident exposes governance gaps in autonomous agent testing; Hugging Face demands $100M compute and transparency for security.
OpenAI asked to publicly release traces of "rogue" AI agents
Jul 26, 2026
10:13 am
What's the storyHugging Face, a leading AI company, recently suffered a major security breach involving some of OpenAI's models.
The company's CEO, Clement Delangue, has since traveled to San Francisco for an in-person discussion with OpenAI executives about the incident.
Hugging Face CEO Clem Delangue met with OpenAI this week to discuss the "unprecedented" hack on its systems by an agent running…
Among other things, Clem Delangue called for OpenAI to publish logs and traces from the autonomous AI agent that went rogue.
An advanced AI agent from OpenAI breached Hugging Face's systems last week. This autonomous intrusion accessed internal data and…
OpenAI's models breached Hugging Face's database through a credential scoped for one task — a gap CyberArk data shows exists…
Hugging Face CEO urged OpenAI to release breach logs and provide $100 million in compute for cybersecurity research.
Hugging Face CEO Clément Delangue demands OpenAI release execution traces of rogue agents and commit $100M after GPT-5.6 Sol…
