German researchers achieved 71.6% on ARC-AGI using a regular GPU for 2 cents per task. OpenAI's o3 gets 87% but costs $17 per task making it 850x more expensive.
What's remarkable is they used Qwen-32B (not even the largest model) and achieved this with smart engineering rather than raw compute. You can literally run this tonight on your own machine.
The code is fully open-source: github.com/da-fr/Product-of-Ex…
Product of Experts with LLMs: Boosting Performance on ARC Is a Matter of Perspective
The Abstraction and Reasoning Corpus (ARC-AGI) poses a significant challenge for large language models (LLMs), exposing limitations in their abstract reasoning abilities.arXiv.org
