Preprint
Jul 2026
Desktop-Delta Bench: Do Computer-Use Models Understand Desktop GUI Transitions?
Desktop-Delta Bench is introduced, an offline step-level benchmark with 2,013 human-verified instances from novel, multi-app Linux trajectories across ~15 applications and 50 task domains that complements end-to-end benchmarks by filling the missing diagnostic layer between GUI grounding and final task success, enabling targeted improvements to desktop CUA verification, reliability, and recovery.
Abhishek Pillai, Samir Kumar Nayak, Yuan-Fan Chen
· 0 citations