A2-WLA

Action model with a latent world model from SJTU’s AutoLab (School of AI) and Anyverse Dynamics, the policy inside the A2-MEAS multi-agent system described in “Toward the Era of Agentic Embodied Intelligence”. The full A2-MEAS system reports a higher RoboTwin 2.0 average (95.05) than the policy alone.

Shanghai Jiao Tong University & Anyverse Dynamics · Nov 6, 2026 · VLAReport an error
1Benchmarks reported
Nov 6, 2026Released
OpenWeights
VLAType

Benchmark results1

RoboTwin 2.0Near saturationSimulation
  • 94.0%, self-reported by the authors, not a leaderboard entryPaper / model card ↗Nov 2026· self-reported· n = 10000· non-standard
94.0%, self-reported by the authors, not a leaderboard entrynot ranked
Not reported on 14 benchmarks
Wrong or missing number? Report an error