Hugging Face used GLM 5.2 to analyze OpenAI security incident

Hugging Face utilized the open-weight GLM 5.2 model to investigate a rogue test model attack after Claude and GPT-4 declined the request. The analysis was completed in several hours.