Submit yours

ReplayARC-AGI-1 · Fresh · GPT-6 Luna, high

Prime Agent, asked to store

Asked after every verdict to store what would help next time.

Held-out tasks, budget used
20.6 solved 96%
Model spend
$1.08 1,022 calls
Instances
100 3.5 hours

Budget used is the share of the cap of 16 that an instance was charged, averaged over the instances: a solved instance costs 1 for each wrong answer and each demonstration, an unsolved one all 16, so lower is better.

Loading the run