# Top ARC-ΑGI-3 scores on Kaggle just went from 7% to 56% [N] — retained record

> Retained reddit record from reddit:MachineLearning.

Canonical: https://highsignal.app/data/records/0e4c5ef9d2fbed49

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:MachineLearning)  published 2026-10-04 10:24 ingested 2026-10-05 02:38

[original source ↗](https://www.reddit.com/r/MachineLearning/comments/1wxcd4k/top_arc%CE%B1gi3_scores_on_kaggle_just_went_from_7_to/)

## retained text



```
https://preview.redd.it/gkjgii48gfth1.png?width=575&format=png&auto=webp&s=4e1c35bdacf41d18a6409baedcddc9432ec79814

This happened over the past 30 days. So smallish local models (Kagglers can only use those), in a harness, just started beating *average* humans at a benchmark intentionally designed to show human superiority.

I'm curious what people here think.

(The leaderboard graphic is a little out-of-date)
```



document fetched 2026-10-05 00:26 · canonical https://reddit.com/r/MachineLearning/comments/1wxcd4k/top_arc%CE%B1gi3_scores_on_kaggle_just_went_from_7_to
