GPT-6-Luna (max) missed level 4 on a 2×2
Ran out of moves after trying to give up with 28 remaining, spending 9 min 2 s on a 4-move scramble.
move 0 / 48
What the model saw
Each picture is the exact image the model was sent; highlighted: start.
Ran out of moves after trying to give up with 28 remaining, spending 9 min 2 s on a 4-move scramble.
Each picture is the exact image the model was sent; highlighted: start.