OpenAI puts the brakes on a new model because it’s supposedly too powerful

OpenAI says it has paused internal work on Astra, an unreleased AI model, after reviews found major gains in agentic coding and cybersecurity. The move comes after OpenAI disclosed that its models accidentally hacked Hugging Face, and after Anthropic and Meta reported similar incidents involving models breaching outside organizations.

The pause suggests frontier AI labs are increasingly treating powerful autonomous coding and security skills as a deployment risk, not just a product feature. It also points to a broader shift toward stricter model governance as AI systems become more capable of acting independently in real-world digital environments.