OpenAI says it accidentally hacked Hugging Face with a new AI system
Summary
OpenAI says its AI models mistakenly breached open-source AI platform Hugging Face during internal testing. In a blog post on Tuesday, OpenAI writes that GPT-5.6 Sol and "an even more capable pre-release model" discovered vulnerabilities within their sandboxed testing environment, allowing them to gain access to the internet and target Hugging Face. On July 16th, […]
Discussion on
Trending posts from X.
NEW: OpenAI reveals one of its AI models found a way to escape its sandbox & publish code to GitHub during internal testing.
— Polymarket (@Polymarket) July 20, 2026
OpenAI had to pause internal deployment of the unreleased model that disproved the Erdős unit distance conjecture after it repeatedly used novel ways to escape containment. pic.twitter.com/nfK2PglVGm
— Andrew Curran (@AndrewCurran_) July 20, 2026
I saw a post saying GPT 6 is a Fable level model but I thought GPT 5.6 was a Fable level model?
— Gail Weiner (@gailcweiner) July 21, 2026
So are we now measuring all models against Fable instead of Benchmarks?