ResWAM
Residual world-action model from Michigan State University, Cisco, the University of Michigan and IBM Research that predicts the visual change between frames (the residual) instead of the full future frame, with no robot pretraining.
Michigan State University et al. · Nov 14, 2026 · World-actionReport an error
82.6%LIBERO+ · #19 of 37
2.3BParameters
1Benchmarks reported
Nov 14, 2026Released
Benchmark results1
LIBERO-PlusSimulation
- Paper / model card ↗Nov 2026
#19 of 3782.6%, self-reported by the authors, not a leaderboard entry
Not reported on 14 benchmarks
Wrong or missing number? Report an error