Continual-ARCby unify

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

Hermes Agent, asked to store

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

Held-out tasks, budget used
21.9 solved 95%
Model spend
$2.87 3,681 calls
Instances
100 10.4 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