Continual-ARC

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

Prime 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
29.6 solved 80%
Epilogue, budget used
39.5 solved 60%
Model spend
$1.92 3,206 calls
Instances
225 4.1 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