★ Top story · ModelsAug 7
Responding to the next frontier of critical cyber capabilities
OpenAI has paused development on its upcoming Astra model to implement additional security controls after internal testing revealed the system possesses advanced cybersecurity capabilities. The company is currently conducting preliminary evaluations to assess how these tools might be misused for offensive tasks. By prioritizing these safeguards now, OpenAI aims to mitigate potential risks before the model is released to the public. This shift reflects a broader industry focus on preventing highly capable artificial intelligence from lowering the barrier to entry for complex cyberattacks.