MemGUI-Bench
MemGUI-Bench: Benchmarking Memory of Mobile GUI Agents in Dynamic Environments
Guangyi Liu, Pengxiang Zhao, Yaozhen Liang, Qinyi Luo, Shunye Tang, Yuxiang Chai, et al. · 2026-02-03
- Modality
- GUI & computer use
- Data
- Real data
- Task · ground truth
- Act. · State+Act.
- Scope
- Memory-specific benchmark
What it tests
- CECross-modal evidencePartially evaluated
- LELong-range evidenceDirectly evaluated
- SCState changesDirectly evaluated
- CFConflicting memoriesNot evaluated
- UQUnanswerable queriesNot evaluated
- MAMemory-grounded actionDirectly evaluated
- GSPersistent interface stateDirectly evaluated
- AOAction–outcome associationDirectly evaluated
- XRCross-session reuseDirectly evaluated
Survey notes
Mobile tasks measure within-session retention and reuse over repeated or mirrored interactions, with memory-fidelity evaluation.