TY  - RPRT
TI  - FailBench: How Reliable are VLMs at Judging Robot Task Success?
AU  - Zaruhi Navasardyan
AU  - Tatul Danielyan
AU  - Hrant Davtyan
PY  - 2026
UR  - https://arxiv.org/abs/2609.03611
ID  - 2609.03611
ER  -