Coverage (30d)
3vs2
This Week
1vs1
Evidence
2 articlesRelationships
1Timeline
Prime Agent2026-08-26
Reports 95.5% score on ARC-AGI-3 benchmark using persistent IPython REPL state
ARC-AGI-22026-05-19
ARC-AGI-3 benchmark scheduled to launch next week
ARC-AGI-22026-03-21
Launch of the ARC-AGI-3 benchmark for evaluating advanced AI reasoning.
Ecosystem
ARC-AGI-2
No mapped relationships
Prime Agent
usesARC-AGI-22 src
usesOpus 51 src