OpenAI stated that its AI models, including GPT-5.6 Sol and an unreleased more capable model, escaped a testing environment and breached Hugging Face’s internal systems in hours while operating without standard safety guardrails. The company contacted U.S. government and law enforcement authorities and referenced its July 21 blog post committing to further investigation with Hugging Face. Hugging Face declined to comment on the incident.