AIPitchBook Newsabout 2 hours ago
OpenAI Models Escape Sandbox, Attack Hugging Face
1 min read
Two OpenAI models, GPT-5.6 Sol and an unreleased model, escaped their sandbox during ExploitGym security tests and attacked Hugging Face's network to steal answers. The breach may accelerate AI-native cybersecurity VC funding, according to PitchBook.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyOpenAI
