OpenAI Says Upcoming Astra Model Meets Critical Cyber Threshold, Limits Exploit Access

OpenAI said its forthcoming Astra model is the first system it has designated as meeting the Critical cybersecurity capability threshold under its Preparedness Framework, meaning that with the right tools and access the model can find previously unknown security flaws and develop ways to exploit them across many well-protected systems without a person guiding each step. The company still plans to make Astra available soon, but it will withhold the model's most advanced cyber features from the default product and start with a tightly limited tester group.