Path to Astra: critical capabilities and frontier safeguards

OpenAI has determined that its Astra model has reached a "Critical" level of cybersecurity capability, meaning it can find previously unknown security flaws and develop ways to exploit them across well-protected systems without human guidance. This designation is significant because it is the first model OpenAI has classified at this level, requiring stronger safeguards during development and before release. The company delayed parts of Astra's development to strengthen protections against misuse, including training the model to refuse harmful requests and implementing monitoring to detect unauthorized activity. Astra will be released with limited access to its most advanced cybersecurity capabilities, initially available only to a group of testers.