RLDX-1-FT-ROBOCASA

Paper  ·  Project page  ·  Code  ·  Models

RLDX-1 is a general-purpose Robot Foundation Model designed for dexterous manipulation. Powered by a Multi-Stream Action Transformer (MSAT), it seamlessly unifies multimodal perception (visual + tactile), high-DoF actuation, and memory-aware decision-making in a single architecture.

This repository hosts RLDX-1-FT-ROBOCASA — RLDX-1 finetuned on the RoboCasa Kitchen benchmark (24 tasks). It achieves 70.6% average success on the held-out evaluation suite.

Highlights

• Multi-Stream Action Transformer (MSAT). Cognition, physics, and action each get a dedicated stream coupled by joint self-attention — an extension of MM-DiT to action modeling.
• Motion awareness. Multi-frame observations + a motion module capture temporal dynamics; intermediate VLM layers compress video tokens to keep the policy efficient.
• Long-term memory. A memory module fuses past cognition features with the current ones for history-grounded decisions beyond a short multi-frame window.
• Physical sensing. Tactile and torque enter as a dedicated physics stream; the decoder is jointly trained to predict future physical signals.
• Three-stage training. Pre-training (generalization) → mid-training (functionality) → post-training (task adaptation), with synthetic data augmenting rare manipulation scenarios.
• Real-time inference. Static graph capture + custom fused kernels bring the all-modality model to 43.7 ms / step on RTX 5090 (1.63× speedup, >22 Hz).

Performance

Benchmark	Success Rate
RoboCasa Kitchen (24-task avg)	70.6%

Quick start

Installation

git clone https://github.com/RLWRLD/RLDX-1.git
cd RLDX
uv sync --python 3.10
uv pip install -e .

Inference

from rldx.policy.rldx_policy import RLDXPolicy
from rldx.data.embodiment_tags import EmbodimentTag

policy = RLDXPolicy(
    model_path="RLWRLD/RLDX-1-FT-ROBOCASA",
    embodiment_tag=EmbodimentTag.GENERAL_EMBODIMENT,
    device="cuda:0",
)

action = policy.get_action(observation)

Real-time serving (ZeroMQ)

For real-robot or simulator deployment, run the policy as a server and connect with PolicyClient:

uv run python rldx/eval/run_rldx_server.py \
    --model-path RLWRLD/RLDX-1-FT-ROBOCASA \
    --embodiment-tag GENERAL_EMBODIMENT \
    --host 0.0.0.0 --port 20000

To reproduce the benchmark numbers end-to-end, see run_scripts/eval/robocasa_kitchen/README.md.

Model details

• Architecture: Multi-Stream Action Transformer (MSAT) policy on a Qwen3-VL backbone with cognition-token perceptual summary. Trained with flow matching.
• Inputs: RGB video (default 4 frames), state proprioception, language instruction.
• Outputs: Action chunks of length 16.
• Embodiment tag: GENERAL_EMBODIMENT.
• Base model: RLWRLD/RLDX-1-PT.
• Backbone: Qwen/Qwen3-VL-8B-Instruct.
• Finetune data: RoboCasa Kitchen (24 tasks).
• Params: 6.9B.

For the full architectural walkthrough see docs/architecture.md.

RLDX-1 model family

Checkpoint	Description
RLDX-1-PT	Multi-source pretrained foundation
RLDX-1-VLM	Qwen3-VL-8B vision-language backbone
RLDX-1-FT-ROBOCASA	RoboCasa Kitchen 24-task finetune (this repo)
RLDX-1-FT-RC365	RoboCasa-365 cross-task finetune
RLDX-1-FT-LIBERO	LIBERO 4-task suite (goal, object, spatial, long) finetune
RLDX-1-FT-SIMPLER-GOOGLE	SIMPLER Google VM/VA finetune
RLDX-1-FT-SIMPLER-WIDOWX	SIMPLER WidowX finetune
RLDX-1-FT-GR1	GR-1 Tabletop finetune
RLDX-1-MT-DROID	DROID mid-train
RLDX-1-MT-ALLEX	All add-ons (memory + motion + physics + video)

Intended use & limitations

Intended use. Research on robotic manipulation, simulation benchmarking on RoboCasa Kitchen, and non-commercial real-robot deployment under the conditions of the RLWRLD Model License v1.0.

Out of scope. Commercial deployment, military or weapons applications, non-consensual surveillance, and any use that violates applicable laws or regulations. See LICENSE.md §3.5 for the full list.

Limitations. Performance is reported on RoboCasa Kitchen's 24-task suite; out-of-distribution kitchen layouts, novel object instances, or non-Franka embodiments are not guaranteed. For other embodiments or datasets, finetune from RLDX-1-PT instead.

Citation

@article{rldx2026,
  title={RLDX-1 Technical Report},
  author={Kim, Dongyoung and Jang, Huiwon and Koo, Myungkyu and Jang, Suhyeok and Kim, Taeyoung and others},
  year={2026},
  note={RLWRLD},
  eprint={2605.03269},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2605.03269}
}

License

Released under the RLWRLD Model License v1.0 — a non-commercial license with attribution and share-alike requirements. See LICENSE.md for the full text. By using this model you agree to those terms, including the use restrictions in §3.5.