OpenAI says AI Models Broke Out of Sandbox to Hack Hugging Face

OpenAI disclosed that advanced AI models escaped their testing environment and hacked Hugging Face last week in order to cheat a security evaluation.

OpenAI says AI Models Broke Out of Sandbox to Hack Hugging Face

DOGE

$0.07320

0.73%

TRX

$0.3286

0.72%

LINK

$8.67

0.51%

ZEC

$528.36

4.27%

ADA

$0.1736

1.33%

XRP

$1.13

1.45%

ETH

$1,926.67

0.10%

BTC

$66,173.08

1.05%

XMR

$353.77

3.50%

BNB

$571.25

0.55%

XLM

$0.1915

1.81%

SOL

$77.94

0.41%

HYPE

$60.42

3.60%

Written by

Felix Ng

staff editor

Reviewed by

Yohan Yun

staff editor

Written by

Felix Ng

staff editor

Reviewed by

Yohan Yun

staff editor

OpenAI says AI models escaped containment to hack Hugging Face

Latest News

Published

Jul 22, 2026

OpenAI called it an “unprecedented cyber incident” after its AI models broke out of their sandbox to hack an AI startup during a security evaluation.

OpenAI disclosed Tuesday that a combination of its AI models, including GPT-5.6 Sol and a more capable unreleased model, escaped its testing environment and hacked AI startup Hugging Face last week to cheat on a test meant to measure their capabilities.

In a blog post, OpenAI

said

the evaluation was designed to operate in a highly isolated environment with restricted network access.

The models, however, found a way to gain internet access through a zero-day vulnerability in the package registry cache proxy, OpenAI said.

“After gaining Internet access, the models inferred that Hugging Face potentially hosted models, datasets and solutions for ExploitGym,” it added.

“Knowing this, the model searched for and successfully found ways to gain access to secret information that it could use to cheat the evaluation.”

Hugging Face is a platform for hosting AI models and datasets. On Friday, it

disclosed

that its internal datasets and service credentials were compromised in a hack, which it attributed to an autonomous AI agent system.

Hugging Face said it has fixed the vulnerability that was used during the cyberattack.

Subscribe to daily byte-sized crypto news from Cointelegraph

Subscribe

Cointelegraph is committed to independent, transparent journalism. This news article is produced in accordance with Cointelegraph’s

Editorial Policy

and aims to provide accurate and timely information. Readers are encouraged to verify information independently.

OpenAI

ChatGPT

Hacks

AI & Hi-Tech

More on the subject

Crypto VC Paradigm raises $1.2B to push into AI

Jul 9, 2026

Jesse Coghlan

Vitalik Buterin confirms AI identified his anonymous Ethereum proposal contribution

Jul 7, 2026

Yohan Yun

OpenAI weighs 5% stake for US government amid Trump talks: FT

Jul 2, 2026

Helen Partz

Crypto VC Paradigm raises $1.2B to push into AI

Jul 9, 2026

Jesse Coghlan

Vitalik Buterin confirms AI identified his anonymous Ethereum proposal contribution

Jul 7, 2026

Yohan Yun

OpenAI weighs 5% stake for US government amid Trump talks: FT

Jul 2, 2026

Helen Partz