descriptive claim
ARC-AGI's quality check runs a random agent for one million steps, and the agent does not beat the Locksmith benchmark.
Unresolved speaker · ⚡️ARC-AGI-3: The Interactive Reasoning Benchmark · 18 Jul 2025
In their words
The transcript segments this claim was taken from
“And one of the quality checks that we do is we run a random agent at a million steps to see if it beats it or not. And no,”
“it doesn't beat lockstep at all or locksmith. And so what we'll do is when we come out with Arc AGI.”18:27Report a misattribution
0.97how strongly it was put, not whether it is true
0.68importance to the conversation
0.95confidence the claim was read correctly