MolmoSpaces: A Large-Scale Open Ecosystem for Robot Navigation and Manipulation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917280789561344 |
|---|---|
| author | Kim, Yejin Pumacay, Wilbert Rayyan, Omar Argus, Max Han, Winson VanderBilt, Eli Salvador, Jordi Deshpande, Abhay Hendrix, Rose Jauhri, Snehal Liu, Shuo Shafiullah, Nur Muhammad Mahi Guru, Maya Eftekhar, Ainaz Farley, Karen Clay, Donovan Duan, Jiafei Guru, Arjun Wolters, Piper Herrasti, Alvaro Lee, Ying-Chun Chalvatzaki, Georgia Cui, Yuchen Farhadi, Ali Fox, Dieter Krishna, Ranjay |
| author_facet | Kim, Yejin Pumacay, Wilbert Rayyan, Omar Argus, Max Han, Winson VanderBilt, Eli Salvador, Jordi Deshpande, Abhay Hendrix, Rose Jauhri, Snehal Liu, Shuo Shafiullah, Nur Muhammad Mahi Guru, Maya Eftekhar, Ainaz Farley, Karen Clay, Donovan Duan, Jiafei Guru, Arjun Wolters, Piper Herrasti, Alvaro Lee, Ying-Chun Chalvatzaki, Georgia Cui, Yuchen Farhadi, Ali Fox, Dieter Krishna, Ranjay |
| contents | Deploying robots at scale demands robustness to the long tail of everyday situations. The countless variations in scene layout, object geometry, and task specifications that characterize real environments are vast and underrepresented in existing robot benchmarks. Measuring this level of generalization requires infrastructure at a scale and diversity that physical evaluation alone cannot provide. We introduce MolmoSpaces, a fully open ecosystem to support large-scale benchmarking of robot policies. MolmoSpaces consists of over 230k diverse indoor environments, ranging from handcrafted household scenes to procedurally generated multiroom houses, populated with 130k richly annotated object assets, including 48k manipulable objects with 42M stable grasps. Crucially, these environments are simulator-agnostic, supporting popular options such as MuJoCo, Isaac, and ManiSkill. The ecosystem supports the full spectrum of embodied tasks: static and mobile manipulation, navigation, and multiroom long-horizon tasks requiring coordinated perception, planning, and interaction across entire indoor environments. We also design MolmoSpaces-Bench, a benchmark suite of 8 tasks in which robots interact with our diverse scenes and richly annotated objects. Our experiments show MolmoSpaces-Bench exhibits strong sim-to-real correlation (R = 0.96, \r{ho} = 0.98), confirm newer and stronger zero-shot policies outperform earlier versions in our benchmarks, and identify key sensitivities to prompt phrasing, initial joint positions, and camera occlusion. Through MolmoSpaces and its open-source assets and tooling, we provide a foundation for scalable data generation, policy training, and benchmark creation for robot learning research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_11337 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | MolmoSpaces: A Large-Scale Open Ecosystem for Robot Navigation and Manipulation Kim, Yejin Pumacay, Wilbert Rayyan, Omar Argus, Max Han, Winson VanderBilt, Eli Salvador, Jordi Deshpande, Abhay Hendrix, Rose Jauhri, Snehal Liu, Shuo Shafiullah, Nur Muhammad Mahi Guru, Maya Eftekhar, Ainaz Farley, Karen Clay, Donovan Duan, Jiafei Guru, Arjun Wolters, Piper Herrasti, Alvaro Lee, Ying-Chun Chalvatzaki, Georgia Cui, Yuchen Farhadi, Ali Fox, Dieter Krishna, Ranjay Robotics Artificial Intelligence Computer Vision and Pattern Recognition Deploying robots at scale demands robustness to the long tail of everyday situations. The countless variations in scene layout, object geometry, and task specifications that characterize real environments are vast and underrepresented in existing robot benchmarks. Measuring this level of generalization requires infrastructure at a scale and diversity that physical evaluation alone cannot provide. We introduce MolmoSpaces, a fully open ecosystem to support large-scale benchmarking of robot policies. MolmoSpaces consists of over 230k diverse indoor environments, ranging from handcrafted household scenes to procedurally generated multiroom houses, populated with 130k richly annotated object assets, including 48k manipulable objects with 42M stable grasps. Crucially, these environments are simulator-agnostic, supporting popular options such as MuJoCo, Isaac, and ManiSkill. The ecosystem supports the full spectrum of embodied tasks: static and mobile manipulation, navigation, and multiroom long-horizon tasks requiring coordinated perception, planning, and interaction across entire indoor environments. We also design MolmoSpaces-Bench, a benchmark suite of 8 tasks in which robots interact with our diverse scenes and richly annotated objects. Our experiments show MolmoSpaces-Bench exhibits strong sim-to-real correlation (R = 0.96, \r{ho} = 0.98), confirm newer and stronger zero-shot policies outperform earlier versions in our benchmarks, and identify key sensitivities to prompt phrasing, initial joint positions, and camera occlusion. Through MolmoSpaces and its open-source assets and tooling, we provide a foundation for scalable data generation, policy training, and benchmark creation for robot learning research. |
| title | MolmoSpaces: A Large-Scale Open Ecosystem for Robot Navigation and Manipulation |
| topic | Robotics Artificial Intelligence Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.11337 |