Public KG record
ARC-AGI-3
Recorded overview
A benchmark using 2D puzzle games to measure AI agent's ability to learn, reason
An evidence-backed entity from the project knowledge graph; detailed tool metadata is still being added.
Evidence coverage
The KG currently records 4 known information dimensions and has 13 undocumented dimensions, linked to 9 public source articles. These counts describe record coverage, not model performance or verified fit.
Related public records
- Claude Code — evaluated on
- Prime Agent — applied to
- Compaction — improves
- Retained reasoning — improves
- Claude Opus 5 — evaluated on