← Leaderboard

PatchWAM

World-action
PatchWAM Team · Sep 2026

World-action model built on an image-generation backbone, listed on both RoboTwin 2.0 and RoboCasa365.

Type
World-action
Parameters
—
Backbone
FLUX.2-klein-base-4B + Qwen3-4B
Weights
Open
License
–
Commercial use
Unclear
Released
Sep 2026
Robots
Aloha-AgileX (sim)
SO-101
–
LeRobot
–

Benchmark results

6 of 13 benchmarks
#13 of 2647.6%
#19 of 4023.6%, self-reported by the authors, not a leaderboard entry
#16 of 2480.2%, self-reported by the authors, not a leaderboard entry
#4 of 657.9%, self-reported by the authors, not a leaderboard entry
RoboTwin 2.0Near saturationSimulation
#8 of 2379.1%
LIBEROSaturatedSimulation
#20 of 4698.4%, self-reported by the authors, not a leaderboard entry
Not reported on 7 benchmarks
Robustness drop: 98.4% on LIBERO → 80.2% on LIBERO-Plus → 57.9% on LIBERO-PRO. The perturbed variants show how much of the LIBERO score is memorization.

Adoption

updated Nov 17, 2026
HF downloads · 30d
–
HF likes
0
GitHub stars
1
GitHub forks
0

Trained on

Wrong or missing number? Report an error