Reasoning benchmark
ARC-AGI-1 leaderboard
ARC-AGI-1 Semi-Private Evaluation. Every model the catalog carries a published ARC-AGI-1 value for, ranked by that value.
CategoryReasoning
MeasureVerified accuracy
TasksSemi-private ARC-AGI-1 evaluation set
DifficultyAbstract visual reasoning
Published byARC Prize leaderboard
ARC Prize fluid-intelligence benchmark using novel visual grid transformations.
ARC-AGI-1 ranking
3 models with a published ARC-AGI-1 value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.
| Rank | Model | Provider | Verified accuracy |
|---|---|---|---|
| 1 |  GPT-6 Astra | OpenAI | 98.50 |
| 2 |  Claude Fable 5.1 | Anthropic | 97.50 |
| 2 | Claude Opus 5 | Anthropic | 97.50 |
Evidence key: Observed
Rows are ordered by the value ARC Prize leaderboard published, highest first. The source does not state whether a higher value is the better result, so this page does not either: for a benchmark that measures a rate of failure, read the table from the bottom.