Hugging Face uses GLM 5.2 to analyze OpenAI security incident

Hugging Face researchers utilized the open-weight GLM 5.2 model to investigate a rogue test model attack after Claude and GPT declined the request. The analysis was completed in hours, highlighting the utility of open-weight models in cybersecurity.