Overview
- GPT-5 has passed OpenAI’s final internal assessments and red-teaming, with an official release expected any day.
- During testing, Altman said he was “scared” by GPT-5’s capabilities and admitted feeling “useless” next to its intelligence.
- He compared the model’s development to the Manhattan Project, signaling concerns over its unprecedented potency.
- Altman criticized current oversight, warning “there are no adults in the room” as AI innovation outpaces governance.
- GPT-5 is expected to deliver a significant leap in speed and reasoning compared to its GPT-4 predecessor.