Hierarchical Reasoning Model (HPM)

Closed weights Sapient Intelligence 27M parameters August 2025

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
Sapient Intelligence
Country
Singapore
Published
4 August 2025
Authors
Guan Wang, Jin Li, Yuhao Sun, Xing Chen, Changling Liu, Yue Wu, Meng Lu, Sen Song, Yasin Abbasi Yadkori

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Language, Vision, Multimodal
Task
Visual puzzles

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters
27M

27M

Training data
tokens

1000 training samples

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights
Closed — provider access only
Model access
Unreleased
Training code
Open source

Apache 2.0 https://github.com/sapientinc/HRM

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Why it is tracked
SOTA improvement

"With only about 1000 training examples, the HRM (~27M parameters) surpasses state-of-the-art CoT models on inductive benchmarks (ARC-AGI) and challenging symbolic tree-search puzzles (Sudoku-Extreme, Maze-Hard) where CoT models failed completely."

Record confidence
Confident

Sources

Where this record came from and when it was last checked.

Reference
Hierarchical Reasoning Model
Last updated
11 February 2026

What the numbers mean

About this model

Hierarchical Reasoning Model (HPM) was published by Sapient Intelligence, in Singapore, in August 2025.

It works in Language, Vision, Multimodal, and is recorded as doing visual puzzles.

This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.

How it was trained

The reason it appears in this catalogue at all is sOTA improvement.

Answers

Hierarchical Reasoning Model (HPM) — common questions

01

When was Hierarchical Reasoning Model (HPM) released?

Hierarchical Reasoning Model (HPM) was published in August 2025.

02

What is Hierarchical Reasoning Model (HPM) used for?

Hierarchical Reasoning Model (HPM) works in Language, Vision, Multimodal, and is recorded as handling visual puzzles. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

03

What GPU do I need to run Hierarchical Reasoning Model (HPM)?

None. Hierarchical Reasoning Model (HPM) is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

04

Is Hierarchical Reasoning Model (HPM) open source?

No. Hierarchical Reasoning Model (HPM) has not had its weights published, so it exists only as a service controlled by its owner.

05

How many parameters does Hierarchical Reasoning Model (HPM) have?

Hierarchical Reasoning Model (HPM) has 27M parameters. 27M. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

06

Who created Hierarchical Reasoning Model (HPM)?

Hierarchical Reasoning Model (HPM) was published by Sapient Intelligence, based in Singapore.

Source

Original publication

Record last updated 11 February 2026

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.