ARC PRIZE 2025
Website: https://www.arcprize.org
Github: https://github.com/arcprize/ARC-AGI-2
LLMs should code like this to verify it's solutions. I am focusing on completing the 120 evaluation tasks first before creating a Kaggle solution, so that LLM labs can use this fully to train their next models.
Setup instructions:
pip install -e .
Run tests:
pytest
Run tasks:
python task_<task_number>.py