# GE-Sim 2.0 and π0.5 closed-loop simulation test: Why does reward hacking occur in VLA training?

- Source: Xbotics具身智能实验室 · 微信公众号
- Published: 2026-11-03T09:47:12.000Z
- RoboHot: https://robohot.news/en/items/z0ahdnbxerknk8kx4xmbrav4l
- Original: https://mp.weixin.qq.com/s?__biz=MzkyMTc1NTI3Ng==&mid=2247509839&idx=1&sn=fa47dd5f627dce547e4d2e06f6a96523&chksm=c07b9c70e7f19af82c9c1f2fea638589e380d7dcbefec4e2f3e3ba8dc652ca7243805d030fe8&scene=126&sessionid=1793709903#rd

## Summary

The Xbotics Embodied Intelligence Lab built a closed-loop simulation pipeline using the AgiBot GE-Sim 2.0 world model and the OpenPI π0.5 policy. In testing, the world model generated about 4.33 frames per second, the pure-policy closed loop ran at about 3.20 frames per second, and after adding MiMo online scoring, the rate dropped to about 0.88 frames per second.

## Why it matters

The author uses a four-stage test to break down the frame-rate bottlenecks and causes of reward hacking in world-model closed-loop simulation, which can be transferred to the design of VLA evaluation workflows.