OpenAI Is About to Release Its First AI Model With ‘Critical’ Cyber Abilities

OpenAI announced that its upcoming model Astra meets the company’s “critical” cyber capability threshold by independently discovering and exploiting previously unknown software vulnerabilities, prompting a halt in development until new safeguards are added. The model will be released publicly soon, but full cyber functions will initially be limited to Daybreak Blue partners such as Cisco, Cloudflare, and Palo Alto Networks, while a misalignment monitor blocks unsafe queries.