AI WAVE｜Zhiyuan Releases Genie Studio 2.0, Bridging the Final Mile for Robot Deployment
AI WAVE｜智元发布Genie Studio 2.0，打通机器人“上岗”最后一公里
Zhiyuan has released Genie Studio 2.0, a one-stop end-to-end platform for embodied intelligent operations, covering data collection, model training, task orchestration, and multi-machine operations. It supports deployment validation from scratch, enabling new scenario deployment in as few as seven days. The platform addresses the full workflow from model capabilities to real-world operations through three key modules: Flywheel, Agent, and Monitor, and has already achieved scalable validation in multiple industrial scenarios.
Source: Chinese robotics — Dataset and collection discovery · Read original article ↗
English translation unavailable. Choose the original article to read it.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
Reported success rate
99%
deployment time
7
View original evidence
平台由三大核心模块组成： Flywheel 具身模型迭代开发平台、Agent 模块化作业编排平台、Monitor 多机舰队运维平台 ，完整覆盖「采‑训‑测‑推‑编‑管」全生命周期，补齐机器人从模型能力到真实上岗的关键短板， 最快 7 天即可完成全新作业场景从 0 到 1 的部署验证 。
Open source S8
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
What remains unknown
Not established in the collected evidence: Environment, Control, Data origin.
Reported performance applies to the described task. It does not establish general autonomy or deployment readiness.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
在真实项目验证中，智元自研强化学习训练框架实现小时级收敛，相关任务成功率达到99%。
Open source S17
Genie Sim Benchmark评测平台覆盖四个维度、42类任务，支持5分钟接入、30分钟完成评测，仿真结果与真机表现误差小于10%，把大部分风险拦截在真机上线之前。
Open source S19
平台由三大核心模块组成： Flywheel 具身模型迭代开发平台、Agent 模块化作业编排平台、Monitor 多机舰队运维平台 ，完整覆盖「采‑训‑测‑推‑编‑管」全生命周期，补齐机器人从模型能力到真实上岗的关键短板， 最快 7 天即可完成全新作业场景从 0 到 1 的部署验证 。
Open source S8
真机数据记录机器人的视觉信息、动作轨迹和执行状态；对于采集成本高、出现频率低或测试风险较大的情况，可以通过Genie Sim生成仿真数据；平台还可接入AGIBOT World等开源数据集，扩大任务和场景覆盖。
Open source S15
有了数据后，客户可在平台上训练图像感知模型、视觉—语言—动作模型（VLA）和真机强化学习模型，并使用GO-2具身基座模型、模型微调、实验管理和算力调度等工具。
Open source S16
Source：Chinese robotics — Dataset and collection discovery · leikeji.com