= ARC-AGI-2
{c}
{title2=2025-03-24}
* https://github.com/arcprize/ARC-AGI-2
* https://arcprize.org/play?task=1ae2feb7
Has the following structure:
* <ARC-AGI-2 training dataset>[Training Set]: 1000 tasks: a variety of difficulty levels, easy to hard, designed to contain all "primitives" needed for eval
* <ARC-AGI-2 public evaluation dataset>[Public Eval Set]: 120 tasks
* Semi-Private Eval: 120 hard tasks, may have been exposed to limited third-parties eg. via API
* Private Eval Set: 120 tasks, never exposed to third parties
Back to article page
