RoboTwin 2.0
Bimanual data generator and policy-generalization benchmark with public trajectories.
Identity and state
- Public code, trajectories, and leaderboard Public openness · reviewed 2026-09-04
Reviewed relationships
| State | Related entity | Relation | Scope | Boundary |
|---|---|---|---|---|
| No reviewed relations are published for this entity yet. | ||||
Canonical evidence claims
Claims are projected from authored tables. A dash remains an explicit missing value; it is never filled from a neighboring product or name match.
| Metric | Value | Class |
|---|---|---|
| Asset library | RoboTwin-OD: 731 objects across 147 categories [benchmarks-r34] | vla-manipulation-eval |
| Claim boundary | 50 benchmark tasks are distinct from the 10-task code-generation study and 4-task real-world study; those method scores stay in the paper [benchmarks-r34] | vla-manipulation-eval |
| Embodiment in the source | dual-arm pairings of Aloha-AgileX, ARX-X5, Piper, Franka, and UR5 manipulators [benchmarks-r34] | vla-manipulation-eval |
| Engine / renderer | — | vla-manipulation-eval |
| Evaluation scope | 5 randomization axes: clutter, lighting, background, tabletop height, and language; clean versus randomized policy evaluation [benchmarks-r34] | vla-manipulation-eval |
| First public date | arXiv 2025-06-22 [benchmarks-r34] | vla-manipulation-eval |
| Kind | scalable bimanual data generator and policy-generalization benchmark [benchmarks-r34] | vla-manipulation-eval |
| Language-conditioned | yes — trajectory-level language is one randomization axis [benchmarks-r34] | vla-manipulation-eval |
| Medium | simulation benchmark and data generator with a separate sim-to-real study [benchmarks-r34] | vla-manipulation-eval |
| Paper | arXiv:2506.18088 [benchmarks-r34] | vla-manipulation-eval |
| Primary metric family | task success under clean and hard randomized conditions on the full 50-task protocol [benchmarks-r34] | vla-manipulation-eval |
| Public code | yes — MIT-licensed official repository [benchmarks-r35] | vla-manipulation-eval |
| Public data | yes — object library and pre-collected trajectories are linked from the official project [benchmarks-r35] | vla-manipulation-eval |
| Public leaderboard | yes — official leaderboard linked from the project [benchmarks-r35] | vla-manipulation-eval |
| Released trajectories | 100,000+ pre-collected expert trajectories across 50 tasks and 5 embodiments [benchmarks-r34] | vla-manipulation-eval |
| Task structure | 50 dual-arm tasks in the paper; the project describes over 50 supported tasks [benchmarks-r34] [benchmarks-r35] | vla-manipulation-eval |
Source ledger
- Research RoboTwin 2.0 — “RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation”. Type: research paper. ID: arXiv:2506.18088. Retrieved: 2026-09-04. reviewed 2026-09-04
- Research RoboTwin 2.0 — official project site with linked code, data, documentation, and leaderboard. Type: research project page. Retrieved: 2026-09-04. reviewed 2026-09-04
What is not inferred
No fuzzy entity matching, adjacent-column borrowing, universal compatibility, commercial maturity, or checkpoint-to-product identity. Absence of a relation means unreviewed or unsupported here—not proven incompatibility.