ARC-AGI-1
Saturated
The original Abstraction and Reasoning Corpus task set from the ARC Prize Foundation, scored on grid-to-grid puzzle solutions.
- Domain
- agent-eval
- Published
- Sep 2026
- Notable for
- The first ARC-AGI generation, now near-saturated by frontier models.
Cite
Notes
Only stored in your browser.
Top score 98.5% by GPT-6 Astra - 1 model reporting (1 frontier)
Top models
1
FAQ
- What is ARC-AGI-1?
- The original Abstraction and Reasoning Corpus task set from the ARC Prize Foundation, scored on grid-to-grid puzzle solutions.
- What is the current top score on ARC-AGI-1?
- The top reported score is 98.5% by GPT-6 Astra, across 1 model reporting (1 from frontier labs).