François Chollet emphasizes ARC-AGI-3 is not an AGI final exam and explains feasibility and purpose, while Lisan al Gaib fixates on scoring, costs, and comparability to ARC-AGI-1/2.
Announcing ARC-AGI-3
The only unsaturated agentic intelligence benchmark in the world
Humans score 100%, AI <1%
ARC-AGI-3 is out now!
evaluate agentic intelligence via interactive reasoning environments
Today we're launching ARC-AGI-3
It is the only unsaturated agent benchmark in the world
AI scores <1%
This gap between human and AI performance proves we do not have AGI
ARC-AGI is *not* a final exam that you pass to claim AGI.
The benchmarks target the residual gap between what's hard for AI and what's easy for humans.
To be clear, all ARC-AGI-3 environments are feasible by humans
Pretty much as expected only GPT-5.4, Opus 4.6 and Gemini 3.1 score above 0
costs are astronomical
This finding is one of many signals tracked across Artificial Intelligence. The live feed updates every few hours with new expert voices, debates, and emerging ideas.
← Back to Artificial Intelligence