arXiv 2610.32965 · cs.RO · 2026-10-23
RynnUnify: Towards Unified Embodied Brain
Minghao Zhu, Bohan Hou, Kun Xiang, Zixuan Wang, Yinzhou Tang, Yifei Yang, Zhikai Wang, Tianyi Zhang 等 23 人
DAMO Academy
一句话要点
RynnUnify 是一个16B统一具身模型,融合视觉语言推理实现长时域规划与物理预测,在多个任务中表现领先,并支持代理决策和机器人控制。
RynnUnify achieves state-of-the-art embodied understanding and planning, with agent and robot control extensions demonstrating strong real-world performance.
要点
- 现有具身模型在长时域视觉规划与物理世界预测上能力不足,缺乏统一的推理框架。
- 通过视频-语言交错推理实现空间推理、物理预测和状态接地,结合强化学习构建代理扩展以预测未观测状态,进一步发展为机器人控制模块。
- 在Rynn-VP任务中优于Cosmos 3 Nano 9.75分,在Rynn-GEU上提升6.0分,机器人任务成功率分别达51.5%和46.7%。
- 为具身理解、视觉规划与控制提供了统一基础,但实际部署仍需进一步验证在复杂动态环境中的鲁棒性。
English overview
RynnUnify presents a 16B unified embodied model that combines video-language reasoning for long-horizon visual planning and physical prediction. It introduces an agent extension via reinforcement learning to reason about unobserved states, and a world-action module for robot control. Evaluated on Rynn-VP and Rynn-GEU benchmarks, it outperforms existing models in embodied understanding and planning, achieving high success rates in real-world robotic tasks.
只根据 arXiv 官方摘要,由开源模型整理,经数字/专名自动核对(必须在摘要中出现),以论文原文为准。
阅读 arXiv 原文 PDF Hugging Face 讨论 代码仓库 项目页
主题与学科
同主题:机器人与具身
- Has GPT-6-Astra Achieved Embodied Intelligence? ▲421
- In-Context Learning for Robots: Methods and Applications ▲388
- GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives ▲310
- HexaShell: A Thin Harness for Exploring the Limits of VLM Agents as Robot Coding Agents ▲294
- Reflex: Visually Grounded Reactive Humanoid Control ▲154
同主题:推理
- LoopVL: Recurrent Visual Intelligence 2026-09-29 · ▲417
- LEV: Thinking Fast and Deep with Looped Decision Models 2026-10-14 · ▲337
- Thinking Inertia: LLMs Keep Thinking When Told Not To 2026-10-08 · ▲241
- Rethinking Latent Visual Reasoning: Grounding Latent Reasoning in Visual Evidence 2026-09-28 · ▲224
- On-Policy or Off-Policy Learning? A Systematic Study of Distillation Dynamics 2026-09-28 · ▲205
数据说明
元数据来自 arXiv API(已核实);点赞来自 Hugging Face 每日论文;引用来自 Semantic Scholar。最近刷新 2026-10-23。
要点由本站根据摘要自写,不转载摘要原文,请阅读原文。