RoboTwin 2.0 leaderboard

Bimanual manipulation benchmark with 50 tasks and strong domain randomization; leaderboard policies train on 50 clean demos per task. Listing requires public code and weights, and top entries exceed 95% on the Easy setting.

Simulation · Aloha-AgileX (bimanual) · SAPIENReport an error
89.0%Best · CloudRobo-WM
33Models ranked
Near saturationSaturation
Nov 18, 2026Numbers checked
#ModelTypeScoreDateSource
1CloudRobo-WMCloudRobo TeamWorld-action89.0%Nov 15, 2026
2Awomo-0.6AwomoWorld-action86.5%Nov 14, 2026
3GO-3 Nano PreviewAGIBOTVLA85.0%Oct 25, 2026
4PrismWAMTsinghua University & NVIDIAWorld-action83.7%Oct 23, 2026
5PICA-WAMShanghai Jiao Tong UniversityWorld-action82.7%Oct 30, 2026
6MWA-PhoenixAnyverse & OpenPhoenixWorld-action79.5%Oct 21, 2026
7SLAHKUST(GZ) & OLAVLA79.2%Nov 14, 2026
8PatchWAMPatchWAM TeamWorld-action79.1%Oct 2, 2026
9ME-Dex-1.0Li AutoVLA78.8%Sep 24, 2026
10EWAMEWAM TeamWorld-action77.1%Oct 29, 2026
11ViGARPKU DAGroup & AGIBOTWorld-action74.5%Oct 11, 2026
12UniWAMHKUST(GZ) & OLAWorld-action71.3%Sep 1, 2026
13GigaBrain-0.7GigaAIVLA67.4%Aug 15, 2026
14WorldLAISCASVLA64.5%Oct 24, 2026
15WorldScape Policy 2.0Manifold AIWorld-action61.7%Sep 24, 2026
164D-WAMOpenHelix RoboticsWorld-action61.6%Aug 18, 2026
17π0.5Physical IntelligenceVLA58.3%Aug 10, 2026
18Spatial ForcingOpenHelix RoboticsVLA52.0%Aug 17, 2026
19X-WAMTsinghua University & Xiaomi RoboticsWorld-action47.9%Aug 10, 2026
20X-VLATsinghua AIR et al.VLA44.5%Aug 10, 2026
21ABot-M0AMAP CV LabVLA43.9%Aug 17, 2026
22EventVLAUSTC et al.VLA40.7%Aug 10, 2026
23Xiaomi-Robotics-0Xiaomi RoboticsVLA40.5%Aug 10, 2026
24Fast-WAMTsinghua University & Galaxea AIWorld-action39.8%Aug 10, 2026
25GalaxeaVLA (G0)Galaxea AIVLA37.7%Aug 17, 2026
26Discrete ForcingOpenHelix TeamVLA35.3%Oct 10, 2026
27AHA-WAMShanghai Jiao Tong University et al.World-action33.8%Aug 10, 2026
28π0Physical IntelligenceVLA31.4%Aug 4, 2025
29DP3 (3D Diffusion Policy)Shanghai Qi Zhi Institute et al.Policy30.1%Aug 4, 2025
30StarVLAStarVLA TeamVLA24.8%Aug 17, 2026
31RDT-1BTsinghua UniversityVLA24.1%Aug 4, 2025
32ACTStanford / UC Berkeley et al.Policy15.7%Aug 4, 2025
33Diffusion PolicyColumbia / Toyota Research InstitutePolicy14.3%Aug 4, 2025

Metric: Mean of Easy (clean) and Hard (domain-randomized) success, 100 trials per task. One result per model: RoboFYI re-run, then official leaderboard, then paper, then third party. Runs that deviate from the official protocol are listed on the model page but not ranked.

Wrong or missing number? Report an error