reasoning
ARC-AGI
Francois Chollet's Abstraction and Reasoning Corpus. Grid-based visual puzzles designed to be trivial for humans but hard for machines that memorised the internet. The most publicised general-intelligence benchmark of 2024-25.
Official leaderboard →Top scores
| # | Model | Score |
|---|---|---|
| 1 | o3 | 87.5% (compute-heavy) |
| 2 | gpt-5 | 32.0% |
| 3 | claude-opus-4-8 | 28.5% |
| 4 | gemini-2-5-pro | 15.0% |
Scores are snapshots from public leaderboards at the time of last update. Follow the source link for the live board.