As of: 2026-11-13 13:49 CET.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/m/rdt2/

# RDT2 (and RDT-1B)

Tsinghua University (TSAIL, thu-ml), Robotics Diffusion Transformer family. Status: current. Type: robotics model. Released: 2025-09.

## How to call it

| Provider | Model id | Endpoint | Docs |
|---|---|---|---|
| Hugging Face | `robotics-diffusion-transformer/RDT2-VQ` |  | https://huggingface.co/robotics-diffusion-transformer/RDT2-VQ |
| Hugging Face (RDT-1B, MIT) | `robotics-diffusion-transformer/rdt-1b` |  | https://huggingface.co/robotics-diffusion-transformer/rdt-1b |
| GitHub |  |  | https://rdt-robotics.github.io/rdt2/ https://github.com/thu-ml/RDT2 |

## Pricing

- Price not published.
- Last checked: 2026-09-29

## Specs

- Input: image, text
- Output: action
- Context window: not published
- Max output: not published
- Open weights: yes
- Licence: apache-2.0

## Where it stands

- Newer robotics models in the registry since its release: 21. The 5 newest:
  - Praxis-1 (2026-09-30) https://postcutoff.com/m/praxis-1/
  - Helix 2.5 (2026-09-17) https://postcutoff.com/m/figure-helix-2-5/
  - UnifoLM-WLA-1.0 (2026-09-10) https://postcutoff.com/m/unifolm-wla-1-0/
  - Skild S1 (Skild Brain) (2026-08-25) https://postcutoff.com/m/skild-s1/
  - Generalist GEN-1.5 (2026-08-19) https://postcutoff.com/m/generalist-gen-1-5/
- Weights are public under a permissive licence (Apache-2.0).

## What stands out

- Zero-shot deployment on unseen embodiments (first): RDT2 (8B, Qwen2.5-VL-7B based, residual-VQ action tokens; RDT2-FM flow-matching variant) trained on 10k+ h of UMI-gripper human manipulation from 100+ scenes; authors call it possibly the first foundation model to deploy zero-shot on unseen embodiments (UR5e, Franka FR3) for simple open-vocabulary tasks. Source: https://huggingface.co/robotics-diffusion-transformer/RDT2-VQ
- Large diffusion foundation model for bimanual manipulation (RDT-1B): RDT-1B (Oct 2024, 1.2B) was billed as the largest diffusion-based foundation model for bimanual manipulation, pretrained on 46 datasets (1M+ episodes) and fine-tuned on a 6K+ episode ALOHA dataset. Source: https://arxiv.org/abs/2410.07864

## Notes

'First' claim is the authors' own hedged wording. HF RDT2-VQ repo created 2025-09-22.


Model page: https://postcutoff.com/m/rdt2/