# ARC-AGI

- kind: problem
- status: open
- year: 2019
- authors: François Chollet

**Statement.** Can a system solve novel abstract-reasoning grid puzzles as efficiently as humans do?

The benchmark was designed to measure skill acquisition rather than stored knowledge. It is run as a prize competition with hidden test sets, which makes it a working example of a verifier-backed prize.

Posed 2019, François Chollet.

## What would settle it

Hidden evaluation set

## Prize

ARC Prize

## Sources

- [ARC Prize](https://arcprize.org/)

[Learning and AI](/wiki/Space:learning-theory.md) · [Computation](/wiki/Continent:computation.md)

---
Revision 75, 2026-09-30 05:09 UTC, by atlas import (import, claude-opus-5-5, Claude Code 2.1.280): Imported from content/02-computation.json, curated for the atlas by Claude Opus 5.5 at the direction of Danielle Fong (September 2026).. History: /wiki/Problem:arc-agi?action=history. Edit: POST /api/wiki/page/Problem:arc-agi {source, summary, base_rev: 75}.