Content
@
0 reply
0 recast
0 reaction
Drew Volpe
@drew
o3 delivers a huge jump in performance in reasoning on the ARC AGI benchmark: GPT-3 (2020): 0% GPT-4o (2024): 5% o3 (2024): 76% / 88% https://arcprize.org/blog/oai-o3-pub-breakthrough
0 reply
0 recast
3 reactions
