OpenAI’s Astra Found Two Security Flaws No Human Had Spotted, Then OpenAI Graded Its Own Safety…

OpenAI’s Astra is the first model it rates “Critical” for cybersecurity. The case that it is safe to release rests largely on numbers…Continue reading on Generative AI »

Source: Generative AI Pub — Published September 5, 2026 at 10:05 UTC — Category: Image AI