150M Model Achieves 29.5% on ARC-AGI

The 150M parameter BDH-CQ model achieved a 29.5% score on ARC-AGI-1 using recurrent latent reasoning, with a very low cost of $0.0007 per task.

2026-08-15 ~ 2026-08-15 · 3 related posts

• New Paper: 150M Model Achieves 29.5% on ARC-AGI at $0.0007 per Task — burny_tech · 2026-08-15
• 150M Recurrent Model Scores 29.5% on ARC-AGI at $0.0007 per Task — juanviera23 · 2026-08-15
• 150M-parameter BDH-CQ breaks ARC-AGI cost-accuracy frontier — moschles · 2026-08-15