SHYRS: A proof of concept for specifying how humans solve ARC-AGI-2 puzzles

ARC-AGI Prize Explainer (Includes ARC-AGI-3!)

Worm Arena

Abstract Port Graphs (APG) is a framework for representing programs as graphs for program synthesis. This repo includes a domain specific representation for ARC-AGI

PUMA – Program Understanding & Meta-learning Architecture Neuroscience-inspired meta-learning system for solving ARC-AGI-2 tasks through RFT-(Relational Frame Theory), symbolic reasoning, neural guidance, and test-time adaptation. #ARC2025

Synalinks ARCAGI2 public benchmark

Official package for "A Neural Affinity Framework for Abstract Reasoning." Includes the validated 9-category ARC taxonomy, pre-computed fine-tuning results, and scripts to verify the Compositional Gap.

Pure program synthesis solver for ARC-AGI-2 — 74/1000=7.4%, no neural networks, no LLMs

A very simplistic approach to solving ARC-AGI-2. Designed to be a neural network built off of primitive functions with special tweaks. This model strives to be as simple and quick as possible as well.

Pure-Python baseline solver for Kaggle ARC Prize 2025 on ARC-AGI-2

LAteNT v2 — A 9-agent neuro-symbolic manifold for zero-shot abstraction. This system replaces hardcoded DSLs with a 64-dimensional Latent Transformation Space, implementing autonomous Bayesian Meta-Learning and online dictionary learning to discover causal laws purely from observation. Pure Inductive Intelligence.