OpenAI's o3 system has reached human level on a test for 'general intelligence'
OpenAI's o3 system got a score of 85% from the previous 55% on the ARC-AGI benchmark. This score is on par with the average human score. This advancement by OpenAI has taken us a step closer to creating artificial general intelligence, or AGI, which is something all major AI research labs are trying to achieve.
Заголовки