Arc
The Abstraction and Reasoning Corpus (ARC) is a benchmark designed to measure human-like general fluid intelligence through grid-based reasoning tasks. It consists of 800 tasks (400 training, 400 evaluation) where each task presents input-output grids that require understanding abstract patterns and transformations. Test-takers must produce exactly correct output grids for all test inputs in a task to solve it, with 3 trials allowed per test input. ARC aims to enable fair comparisons of general intelligence between AI systems and humans using priors designed to be as close as possible to innate human priors.
Progress Over Time
Interactive timeline showing model performance evolution on Arc
State-of-the-art frontier
Open
Proprietary
Arc Leaderboard
1 models • 0 verified
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
1 |  Google | — | 1.0M | $0.10 $0.40 |
Notice missing or incorrect data?
FAQ
Common questions about Arc
The Abstraction and Reasoning Corpus (ARC) is a benchmark designed to measure human-like general fluid intelligence through grid-based reasoning tasks. It consists of 800 tasks (400 training, 400 evaluation) where each task presents input-output grids that require understanding abstract patterns and transformations. Test-takers must produce exactly correct output grids for all test inputs in a task to solve it, with 3 trials allowed per test input. ARC aims to enable fair comparisons of general intelligence between AI systems and humans using priors designed to be as close as possible to innate human priors.
The Arc paper is available at https://arxiv.org/abs/1911.01547. This paper provides detailed information about the benchmark methodology, dataset creation, and evaluation criteria.
The Arc leaderboard ranks 1 AI models based on their performance on this benchmark. Currently, Gemini 2.5 Flash-Lite by Google leads with a score of 0.025. The average score across all models is 0.025.
The highest Arc score is 0.025, achieved by Gemini 2.5 Flash-Lite from Google.
1 models have been evaluated on the Arc benchmark, with 0 verified results and 1 self-reported results.
Arc is categorized under general and reasoning. The benchmark evaluates multimodal models.