# Google and Mercor release PLE benchmark: engineering agents meet 91% of individual constraints but only 41.3% fully pass all deliverables

- Source: 图灵人工智能 · 微信公众号
- Published: 2026-11-04T16:00:00.000Z
- RoboHot: https://robohot.news/en/items/blk82ekmbyup9jpl0e2z7ib06
- Original: https://mp.weixin.qq.com/s?__biz=MzIyMzk1MDE3Nw==&mid=2247680080&idx=6&sn=dc53b0430e71933e7685e31ef6113e75&chksm=e99369e37bf42923a3b985c083fb7dcab1ae8557fa8cff3119619e82d9da50c0365c68c229de&scene=126&sessionid=1793811194#rd

## Summary

Google and Mercor released the Physical-Engineering's Last Exam (PLE) benchmark, which evaluates engineering agents on 85 mechanical, electrical, civil, and structural engineering tasks.