0

Arc Agi 2

ARC-AGI-2: A benchmark measuring abstract reasoning through visual grid puzzles requiring rule inference and generalization.

Domain
agent-eval
Published
Nov 2025

Cite

Notes

Only stored in your browser.

Attribution

Leaderboard scores
OpenAI
Attribution policy →

Top score 95.0% by GPT-6 Astra - 1 model reporting (1 frontier)

Top models

1
Arc Agi 2Bar chart with 1 bar. Highest value: GPT-6 Astra at 95.
1 model

FAQ

What is Arc Agi 2?
ARC-AGI-2: A benchmark measuring abstract reasoning through visual grid puzzles requiring rule inference and generalization.
What is the current top score on Arc Agi 2?
The top reported score is 95.0% by GPT-6 Astra, across 1 model reporting (1 from frontier labs).