VLALagBench: Understanding When and Why VLA Inference Latency Matters

Qi Wen

Borui Xu

Hussein Amro

Hongshi Tan

Hanmin Lin

Chaoyi Ruan

Shanghao Liu

Gang Yang

Yuejiang Liu

Jiafei Duan

Bingsheng He

Jason Clemons

Christos Kozyrakis Stanford

Wenqi Jiang

Preprint, 2026


Abstract

VLALagBench benchmarks how VLA inference latency affects robot task success across tasks, object speeds, execution modes, and policies. An empirical prediction model guides the choice of algorithm configurations for a given inference system.