Robotics & physical AI Model or AI service
RDT-1B studies how robots coordinate both arms
The RDT-1B authors describe a diffusion model that predicts actions for two-arm robots from instructions and observations. [2]
- Model access
- Pretrained RDT-1B weights are published on Hugging Face. [1]
- License
- The model card lists MIT for its published weights. [1]
What the model does
The authors posted the paper in October 2024 and made pretrained weights available. They describe testing the policy on tasks involving bimanual manipulation. [1][2]
Why it matters
Two-arm tasks often require the arms to coordinate, such as holding one object steady while moving another. The research gives developers a defined policy and evaluation setup for studying those movements. [2]
Key dates
- RDT-1B preprint first posted [2]
Ways to use it
Robotics teams can inspect the checkpoint and the tested setup before adapting it to their own cameras, actuators and instructions. [1][2]
Model limitations
Evidence comes from the authors’ research tasks and environments. It provides a basis for testing other platforms and tasks. [2]
Sources and verification
Catalog evidence checked How we verify records
- Primary source robotics-diffusion-transformer/rdt-1b
- Primary source RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation
Related profiles
Robotics & physical AI
Gemini Robotics 2.5
Google DeepMind
The models are intended to interpret a robot's surroundings and support compatible robot tasks.
Robotics & physical AI
NEURA 4NE1
NEURA Robotics
A humanoid platform NEURA presents for development and industrial training, with multiple distinct hardware generations.
Robotics & physical AI
LUMO all-terrain humanoid
Shenzhen Dobot (DOBOT)
LUMO is a humanoid robotics platform whose proposed applications include interaction and service tasks.