fchollet/ARC-AGI
The Abstraction and Reasoning Corpus
This project provides a unique set of visual reasoning puzzles designed to test general fluid intelligence, much like an IQ test. You're given a few example input-output grid pairs that establish a hidden rule, then presented with a new input grid where you must infer and apply the rule to produce the correct output grid. It's for researchers and enthusiasts interested in assessing or developing systems with human-like abstract reasoning abilities.
4,726 stars. No commits in the last 6 months.
Use this if you are a researcher or AI developer working on general intelligence, program synthesis, or psychometrics and need a standardized benchmark to test human-like abstract reasoning.
Not ideal if you're looking for a dataset to train models for conventional machine learning tasks like image classification or natural language processing.
Stars
4,726
Forks
704
Language
JavaScript
License
Apache-2.0
Category
Last pushed
Apr 04, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/rag/fchollet/ARC-AGI"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Related tools
neoneye/arc-notes
My writings about ARC (Abstraction and Reasoning Corpus)
Judahmeek/Significance-Hypothesis-Reasoning-System
SHYRS: A proof of concept for specifying how humans solve ARC-AGI-2 puzzles
lkwbr/agi-pal
Machine learning applied to task assignment within a software organization, as a means of...