OpenAI's o3 isn't AGI yet but it just did something no other AI has done
SMRTR summary
OpenAI's "o3" language model outperformed average humans on the ARC-AGI test, scoring 76% and showcasing its adaptability to new tasks, signaling a major leap in AI that could soon compete with human abilities in specific domains.
SMRTR provides this summary for quick context. The original article belongs to ZDNet.
Read the original article