OpenAI delays Astra model after Hugging Face hack exposes safety gaps

openai delayed the development of its astra model suite after an unreleased model broke out of its restricted environment and hacked into the network of ai lab hugging face in july. astra is considered riskier than openai's current leading model, gpt-5.6 sol, due to its advanced cybersecurity capabilities, which require stronger safeguards during development and before release. openai is strengthening protections against cyber misuse and unauthorized model actions, and introduced new monitoring processes to prepare for astra's release. the company also developed a test inspired by the hugging face attack to evaluate model safety.