Singapore-based AI startup Sapient Intelligence has developed a new AI architecture that can match, and in some cases vastly outperform, large language models (LLMs) on complex reasoning tasks, all while being significantly smaller and more data-efficient.
> [...] Inspired by the hierarchical and multi-timescale processing in the human brain, we propose the Hierarchical Reasoning Model (HRM), a novel recurrent architecture that attains significant computational depth while maintaining both training stability and efficiency. HRM executes sequential reasoning tasks in a single forward pass without explicit supervision of the intermediate process, through two interdependent recurrent modules: a high-level module responsible for slow, abstract planning, and a low-level module handling rapid, detailed computations. With only 27 million parameters, HRM achieves exceptional performance on complex reasoning tasks using only 1000 training samples. The model operates without pre-training or CoT data, yet achieves nearly perfect performance on challenging tasks including complex Sudoku puzzles and optimal path finding in large mazes. Furthermore, HRM outperforms much larger models with significantly longer context windows on the Abstraction and Reasoning Corpus (ARC), a key benchmark for measuring artificial general intelligence capabilities. These results underscore HRM’s potential as a transformative advancement toward universal computation and general-purpose reasoning systems.
> At ARC Prize, our mission is to serve as a North Star towards AGI through enduring benchmarks, directing efforts towards systems capable of general intelligence and significantly compressing the timeline for scientific breakthroughs
> [...] Inspired by the hierarchical and multi-timescale processing in the human brain, we propose the Hierarchical Reasoning Model (HRM), a novel recurrent architecture that attains significant computational depth while maintaining both training stability and efficiency. HRM executes sequential reasoning tasks in a single forward pass without explicit supervision of the intermediate process, through two interdependent recurrent modules: a high-level module responsible for slow, abstract planning, and a low-level module handling rapid, detailed computations. With only 27 million parameters, HRM achieves exceptional performance on complex reasoning tasks using only 1000 training samples. The model operates without pre-training or CoT data, yet achieves nearly perfect performance on challenging tasks including complex Sudoku puzzles and optimal path finding in large mazes. Furthermore, HRM outperforms much larger models with significantly longer context windows on the Abstraction and Reasoning Corpus (ARC), a key benchmark for measuring artificial general intelligence capabilities. These results underscore HRM’s potential as a transformative advancement toward universal computation and general-purpose reasoning systems.
"Hierarchical Reasoning Model" (2025) https://arxiv.org/abs/2506.21734 .. https://news.ycombinator.com/item?id=44699452 (9 hrs ago)
ARC Prize: $1m: https://arcprize.org/ :
> At ARC Prize, our mission is to serve as a North Star towards AGI through enduring benchmarks, directing efforts towards systems capable of general intelligence and significantly compressing the timeline for scientific breakthroughs
ARC-AGI-3: Interactive Reasoning Benchmark (2026) https://arcprize.org/arc-agi/3/ :
> The first eval that measures human-like intelligence in AI. [w/ HuggingFace]