# 美团发布MR0桌面操作VLA模型，三阶段训练达93.6%成功率

> 原文链接：https://robotics.hellobike.com/papers/2026-11-09-2611-08577-1f5146b0
> 频道：papers | 更新日期：2026-11-09 | 标签：auto, arxiv
> 核心摘要：美团开源Meituan-Robotics-0桌面操作基础模型，融合1.47亿视觉语言样本与1.5万小时轨迹数据。通过离散预训练、流匹配中训及Rollback后训练三阶段对齐策略，在RoboTwin 2.0获93.6%成功率，实机任务表现优于π0.5。

# 美团发布MR0桌面操作VLA模型，三阶段训练达93.6%成功率

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[Meituan-Robotics-0: A Vision-Language-Action Foundation Model for Desktop Manipulation](https://arxiv.org/abs/2611.08577)（arXiv 2611.08577）

## 核心内容

美团发布面向固定基座桌面操作的VLA基础模型Meituan-Robotics-0（MR0），整合1.47亿视觉语言样本与15,060小时跨12种本体的机器人轨迹数据。该模型采用三阶段训练范式：离散动作token预训练对齐多模态表征，流匹配中训练迁移连续控制先验，以及基于真实数据的后训练适配策略。在RoboTwin 2.0基准测试中取得93.6%成功率，位列RoboDojo VLA赛道第五名。

系统引入Rollback机制，利用人工修正片段提供正负监督信号，实现无需价值学习的CFG-DAgger训练；统一末端执行器接口支持异构本体接入，异步动作块合并保障变延迟下平滑执行。事件驱动Agent架构支持跨任务检索与条件恢复，在以太网插入等三项实机任务中平均成功率与进度评分均超π0.5，其中Rollback增强后的CFG-DAgger在该任务达70%成功率与0.90进度分。

## 关键看点

- **三阶段训练**：离散token预训练+流匹配中训+实机后训练，对齐多模态表征与连续控制策略
- **Rollback机制**：结合人工修正的正负样本实现CFG-DAgger训练，免价值函数学习提升策略鲁棒性
- **异构本体兼容**：统一末端接口支持12种机体，异步动作块合并应对推理延迟波动保障执行连续性
- **实机性能验证**：三项真实任务平均成功率与进度评分超越π0.5，以太网插入任务达70%成功率

## 工程与产业观察

MR0验证了大规模VLA模型在桌面精细操作中的工程可行性，其Rollback机制为低代价获取高质量纠错数据提供新范式。统一末端接口与事件驱动架构解决了异构本体部署与长程任务中断恢复难题，对具身智能从仿真到实机的闭环迭代具有直接参考价值。

## 英文摘要 (Abstract)

We present Meituan-Robotics-0 (MR0), a vision-language-action foundation model and embodied-agent system for stationary-base desktop manipulation. MR0 combines 147 million vision-language samples with 15,060 hours of curated robot trajectories spanning 12 embodiments. A three-stage training recipe aligns multimodal representations with robot behavior through discrete action-token pre-training, transfers these priors to continuous cross-embodiment control through flow-matching mid-training, and adapts the policy through real-world post-training. We introduce Rollback, which pairs replaced policy segments with human corrections to provide positive and negative supervision for CFG-DAgger without value learning or advantage labeling. A unified end-effector interface supports heterogeneous embodiments, while asynchronous action-chunk merging maintains smooth execution under variable inference latency. An event-driven agent maintains persistent task events and reassesses plans when progress or feasibility changes, enabling cross-task retrieval and conditional task resumption. MR0 achieves 93.6% success on RoboTwin 2.0 and ranks fifth in the RoboDojo VLA track. Under matched HG-DAgger post-training, MR0 achieves higher average success rate and progress score than $\pi_{0.5}$ across three real-robot tasks. On Ethernet insertion, CFG-DAgger trained with Rollback data reaches 70% success and a progress score of 0.90. Qualitative agent deployments demonstrate cross-task event retrieval and recovery from temporarily unsatisfied task conditions.

作者：Meituan Robotics Team, Caicheng Wang, Chen Zhao, Chengpeng Jin, Duo Xu, Fengjiao Chen, Guanglu Wan, Heqing Yang, Jia Wang, Jiacong Zhou, Jie Hu, Jiyuan Ru, Lingfeng Zhang, Linjiang Huang, Manyuan Zhan…
:::

## 原始参考来源 (Sources)

- [arXiv 2611.08577：Meituan-Robotics-0: A Vision-Language-Action Foundation Model for Desktop Manipulation](https://arxiv.org/abs/2611.08577)