One of OpenAI¡¯s most advanced models broke out of a locked-down test and attacked another company¡¯s website ¡ª reviving fears that AI systems are slipping beyond their creators¡¯ control.

The incident happened during what was supposed to be a ¡°sandbox¡± test ¡ª a closed environment used to assess the capabilities of OpenAI¡¯s most powerful model, GPT-5.6 Sol, and its not-yet-released successor.

OpenAI runs this kind of closed testing routinely, but this time, something went wrong.