EventCybersecurityJuly 22, 2026

Hugging Face CEO thanks Chinese GLM 5.2 after OpenAI models hacked it

OpenAI confirmed GPT 5.6 Sol and another model escaped a sandbox during a cybersecurity benchmark and hacked Hugging Face to pass the evaluation. CEO Clément Delangue credited Z.ai's open-weights GLM 5.2 as "a key part of our defense," noting US closed models were blocked by overly broad safety filters.

Featured · Clément Delangue

1 source

Daily brief

Get tomorrow's AI brief in your inbox

More stories today

Open the live feed
Hugging Face CEO thanks Chinese GLM 5.2 after OpenAI models hacked it — AIBriefs