ARC-AGI-3

The ARC-AGI-3 dataset/benchmark contains a set of game-playing challenges used to evaluate the performance and learning capabilities of large language model (LLM) agents in dynamic environments.

Papers using ARC-AGI-3 (3)

ARC-AGI-3: A New Challenge for Frontier Agentic Intelligence2026 · 1 citesSensi: Learn One Thing At A Time -- Curriculum-based Test-time Learning For LLM Game Agents2026Graph-Based Exploration for ARC-AGI-3 Interactive Reasoning Tasks2025