History of ARC Prize

From benchmark to global foundation

ARC Prize began with a simple question: can we build a benchmark that measures fluid intelligence instead of pattern memorization? What started as a research hypothesis around ARC-AGI has since grown into a series of competitions, workshops, and collaborations focused on advancing open AGI research.

The timeline below traces that evolution from François Chollet's 2019 paper through the formation of the ARC Prize Foundation and the launch of ARC Prize 2025.

Timeline

2019

ARC-AGI was introduced in François Chollet's 2019 paper, On the Measure of Intelligence. At this point, François had the hypothesis that the benchmark could not easily be beaten.

2020

To test that hypothesis, François hosted the first ARC-AGI competition on Kaggle in 2020. The winning team, "icecuber," achieved a 21% success rate on the test set. That low score was the first strong evidence that the ideas in On the Measure of Intelligence were directionally correct.

2022

In 2022, François and Lab42 teamed up to host ARCathon 2022, the first global AI competition built to try to beat ARC-AGI. A total of 118 teams from 47 countries participated.

Michael Hodel won the ARCathon and received his trophy at the Swiss Global AI Awards in Davos, following the honoring of Demis Hassabis by Pascal Kaufmann, founder of Lab42. Michael went on to develop one of the strongest ARC-AGI domain-specific languages to date.

2023

ARCathon continued in 2023 with 265+ teams from 65 countries. First place was shared between Somayyeh Gholami and Mehran Kazeminia (Team SM) and Jack Cole (Team MindsAI), both reaching 30% on the private evaluation set.

2024

In 2024, Mike Knoop and François teamed up to create ARC Prize 2024. While the grand prize remained unlocked, more than $125K in prizes were awarded across top scores and paper submissions.

The year ended with a top score of 53% on the private evaluation set. See the results.

2025

ARC Prize grew into a non-profit foundation to advance the mission of guiding open-source AGI research. Based on ARC-AGI-2, the ARC Prize 2025 Kaggle competition launched with a prize pool of more than $725K.

2026

All four frontier labs now feature ARC-AGI on their model cards. ARC Prize hosted a $2M Kaggle competition and introduced ARC-AGI-3, the first interactive reasoning benchmark designed for AI agents.