RLDX-1 Technical Report
Dongyoung Kim, Huiwon Jang, Myungkyu Koo, Suhyeok Jang, Taeyoung Kim, Beomjun Kim, Byungjun Yoon, Changsung Jang, Daewon Choi, Dongsu Han, Donguk Lee, Heeseung Kwon, Hojin Jeon, Jaehyun Kang, Jaekyoung Bae, Jihyuk Lee, Jimin Lee, John Won, Joonwoo Ahn, Junhyeong Park, Junyoung Sung, Kyungmin Lee, Minseong Han, Minsung Yoon, Sejune Joo, Seonil Son, Seungcheol Park, Seunggeun Cho, Seungjun Moon, Seungku Kim, Yonghoon Dong, Yongjin Cho, Youngchan Kim, Chang Hwan Kim, Dohyeon Kim, Heecheol Kim, Heewon Lee, Hensen Ahn, Hyungkyu Ryu, Hyunsoo Choi, Hyunsoo Shin, Jaeheon Jung, Jaewoo Kim, Jinwook Kim, Joochul Chang, Joonsoo Kim, Junghun Park, Jungwoo Park, Junho Cho, Junhyeok Park, Junwon Lee, Kangwook Lee, Kwanghoon Kim, Kyoungwhan Choe, Manoj Bhadu, Nayoung Oh, Sangjun Kim, Sangwoo Kim, Seunghoon Shim, Seunghyun Kim, Seungjun Lee, Seungyup Ka, Sungryol Yang, Wook Jung, Yashu Shukla, Yeonjae Lee, Yeonwoo Bae, Jinwoo Shin
arXiv:2605.03269 · 126▲ · cs.RO, cs.AI, cs.LG
View on arXiv →Premium readers get an interactive explainer for this paper — a figure you can poke at, not just read.
What it is
RLDX-1 is a Vision-Language-Action model for robotic manipulation built on the Multi-Stream Action Transformer (MSAT), which extends the Multi-Modal Diffusion Transformer to route each input modality (vision-language cognition, proprioception, actions, and physical signals like torque and tactile) through a dedicated stream coupled by joint self-attention. It adds a motion module for temporal dynamics, an explicit memory queue for long-horizon tasks, a synthetic data pipeline using video generation with motion-consistency filtering, and inference optimizations (CUDA Graph capture plus custom fused kernels).
Why it matters
Standard VLAs handle scene understanding and language generalization but fail on tasks needing motion awareness, memory, or contact sensing. This work packages those capabilities into one architecture plus a real-time inference path, cutting per-step latency from 71.2 ms to 43.7 ms on an RTX 5090 so the robot's observed state does not drift before actions execute. Mostly relevant to robotics practitioners building deployable manipulation policies.
Practical takeaway
If you are building manipulation policies, watch for modality-specific attention streams that can be masked off when a sensor (e.g. tactile) is absent, and for video-generative synthetic data filtered by replaying predicted actions in a simulator to check motion consistency. Code and weights are released on GitHub and HuggingFace.
Key result
On ALLEX humanoid tasks RLDX-1 reached 86.8% success versus roughly 40% for both pi-0.5 and GR00T N1.6; on the memory-dependent Object-in-Box Selection task it hit 91.7% versus ~30% for both baselines. Caveats: real-robot evals use small trial counts (e.g. 24 trials on Pot-to-Cup), and the baselines were trained for matched steps rather than exhaustively tuned.
Jargon
Subscribe
Get the next issue.
Free. One email a week. Unsubscribe any time: no account, no dark patterns.