Continual-ARC

ReplayARC-AGI-2 · Workstreams · GPT-6 Luna, max

Demo-cache floor

The model alone, shown every demonstration it bought for the task again for free.

Stream, budget used
17.0 solved 90%
Epilogue, budget used
32.3 solved 72%
Model spend
$12.20 867 calls
Instances
225 57.8 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.

Order

Loading the run