XE² Arena launches: one LLM agent, 5 robot bodies, 12 real and 400 sim tasks
shangbinfeng · x · 2026-11-13
Researchers launched XE² Arena, a cross-embodiment × cross-environment platform for evaluating and training agentic robots, to be demoed at CoRL 2026.
- Cross-embodiment: one LLM agent runs on 5 robot bodies — YAM dual arms, xArm, xArm + Wuji Hand 2 five-finger hand, humanoid, mobile manipulator — with more coming.
- Cross-environment: every real run gets an automatic MuJoCo sim replica; you can drag-compare real vs. simulated replays for both eval and post-training.
- Pipeline: XEbench (same task across real bodies) → XEsim (real runs rebuilt in sim) → Robot-RL (verifier success as reward), with trained agents looped back into XEbench.
- Coverage: 12 real tasks across anchor/middle/frontier tiers (cup stacking, T-shirt folding, contact-rich) plus 400 sim tasks for RL.
It follows Claude's (Fable 5.2) strong RoboDojo performance. Paper and code in preparation.
More from Embodied
- Video shows an autonomous parking robot handling the full parking job on its own — DavidWells · 2026-11-13
- BoSS benchmark released: 180K+ stellarators, 3TB of simulation data for generative models — ZhongingAlong · 2026-11-13
- Fly brain hooked up to a synth learns to sing Bad Apple — Bedrovelsen · 2026-11-13
- CoRL takeaways: SuperMimic steals the show, dexterity and tactile sensing improve, human data still a bottleneck — KyleMorgenstein · 2026-11-13
- Outworld Labs launches simulation platform with new physics engine for robot learning — ycombinator · 2026-11-13
- YC-backed Mantle launches AI-native maintenance service for robots with 250+ technicians — ycombinator · 2026-11-13