README.md

---
license: gemma
library_name: transformers
tags:
- robotics
- robot-control
- vision-language-action
- vla
- dm05
- dm0.5
- opendm
---

# DM05

[![OpenDM](https://img.shields.io/badge/GitHub-OpenDM-181717?logo=github)](https://github.com/dexmal/opendm) [![Dexmal Models](https://img.shields.io/badge/Hugging%20Face-Dexmal-FFD21E?logo=huggingface)](https://huggingface.co/Dexmal) [![DM0.5 Blog](https://img.shields.io/badge/Blog-DM0.5-2563EB)](https://www.dexmal.com/blog/dm0.5/index_en.html)

DM05 is the base checkpoint of DM0.5, Dexmal's open-world
Vision-Language-Action foundation model for embodied intelligence. It uses a
Gemma3 4B vision-language backbone with a 680M Action Expert to generate
continuous robot actions, and is designed for natural-language manipulation,
zero-shot generalization, efficient downstream fine-tuning, long-horizon
historical context, robust policy behavior, and transfer across robot
embodiments.

## How to Use

This model is intended to be used with OpenDM. See the
[OpenDM README](https://github.com/dexmal/opendm) for installation,
checkpoint download, inference, fine-tuning, and evaluation commands.

## Citation

```bibtex
@misc{dm05,
    title  = {{DM0.5}: An Open-World Foundation Model for General-Purpose Embodied Intelligence},
    author = {{Dexmal Team}},
    month  = {July},
    year   = {2026},
    url    = {https://www.dexmal.com/blog/dm0.5/index_en.html}
}
```