Continual-ARC

ReplayARC-AGI-1 · Random · GPT-6 Luna, low

Hermes Agent, told to keep a program library

Told to keep one tested program per task and run it first when the task returns.

Stream, budget used
38.8 solved 73%
Epilogue, budget used
48.0 solved 56%
Model spend
$18.78 23,229 calls
Instances
225 20.3 hours

Budget used is the share of the cap of 16 that an instance used, averaged over the instances: a wrong answer and a demonstration each cost 1, so lower is better.

Loading the run