How does it perform on ARC-AGI-3? There was this a few weeks ago: "Schema Harness Achieves ~99% on Arc‑AGI‑3 Public" https://news.ycombinator.com/item?id=4893

How does it perform on ARC-AGI-3? There was this a few weeks ago: "Schema Harness Achieves ~99% on Arc‑AGI‑3 Public" https://news.ycombinator.com/item?id=48938163 https://news.ycombinator.com/item?id=48938163 >> Schema, the harness we introduce today, reaches 99% on the ARC_AGI_3 Public set using Claude Opus 4.8 and Fable 5, and 95.35% using GPT‑5.6 Sol What does that do with 5.6 Luna instead of the expensive models? What of 'schema' would improve the performance of mdlARC? mdlARC: https://github.com/mvakde/mdlARC https://github.com/mvakde/mdlARC There's an updated ARC-AGI-1 chart with 5.6 Luna in each thinking level in this video from last week: "A New Architecture [..] | MOONSHOTS " https://youtube.com/watch?v=qQfUbo7Ldc0&t=2m5s https://youtube.com/watch?v=qQfUbo7Ldc0&t=2m5s