Shifts AI evaluation from static benchmarks to interactive agentic environments requiring fluid adaptation.
March 27, 2026
Original Paper
ARC-AGI-3: A New Challenge for Frontier Agentic Intelligence
arXiv · 2603.24621
AI-generated illustration
The Takeaway
ARC-AGI is the industry standard for measuring generalization; version 3 moves the goalposts to agentic reasoning and planning without explicit instructions, where current frontier models still fail significantly (1% vs 100% human).
From the abstract
We introduce ARC-AGI-3, an interactive benchmark for studying agentic intelligence through novel, abstract, turn-based environments in which agents must explore, infer goals, build internal models of environment dynamics, and plan effective action sequences without explicit instructions. Like its predecessors ARC-AGI-1 and 2, ARC-AGI-3 focuses entirely on evaluating fluid adaptive efficiency on novel tasks, while avoiding language and external knowledge. ARC-AGI-3 environments only leverage Core