ARC-AGI: A new leaderboard to evaluate artificial general intelligence

The ARC-AGI competition releases a leaderboard of AI models capable of solving abstract problems.

Published 15sem1 sourceNotableupdated 10sem

≈ 18sSpeed

The fact

Launched by François Chollet, it aims to measure general intelligence rather than specialized performance.

The results could shape future AI research directions.

Click the link to read an article on the topic:

Explore this topic

Auto-synthesis from 1 media source · identified on July 25, 2026