The story’s score and claims

Comments on No. 66 of 106 · mvakde.github.io · edition of Thursday 12 November 2026

37% on ARC-AGI 2 without an LLM

Summary

A solo researcher's mdlARC, a 100M-parameter autoregressive transformer trained from scratch only on ARC puzzles with no LLM, reports 37.08% on the Kaggle public score for ARC-AGI-2, against 27.64% for last year's LLM-based winner.

Comments

Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules