← Leaderboard

EnerVerse

World-action
AGIBOT et al. · Jan 3, 2025

Generative robotics foundation model from AGIBOT with SJTU, Shanghai AI Lab and CUHK that predicts future embodied space with chunk-wise autoregressive video diffusion and decodes actions with a policy head (EnerVerse-A). It reports LIBERO and CALVIN results from a single static camera; an action-conditioned variant, EnerVerse-AC, is released separately as a simulator.

Type
World-action
Parameters
—
Backbone
–
Weights
Unknown
License
–
Commercial use
Unclear
Released
Jan 3, 2025
Robots
Franka Panda (sim)
SO-101
–
LeRobot
–
Inputs
RGB · L

Benchmark results

2 of 13 benchmarks
CALVIN ABC→DNear saturationSimulation
#22 of 243.00 / 5, self-reported by the authors, not a leaderboard entry
LIBEROSaturatedSimulation
#57 of 7088.5%, self-reported by the authors, not a leaderboard entry
Not reported on 11 benchmarks
Wrong or missing number? Report an error