WorldMemArena
WorldMemArena: Evaluating Multimodal Agent Memory Through Action-World Interaction
Chengzhi Liu, Yuzhe Yang, Sophia Xiao Pu, Yepeng Liu, Lin Long, Yichen Guo, et al. · 2026-06-01
- Modality
- Embodied & 3D
- Data
- Hybrid (real + synthetic/simulated)
- Task · ground truth
- QA/Act./Learn · Ev.+State+Act.
- Scope
- Memory-specific benchmark
What it tests
- CECross-modal evidenceDirectly evaluated
- LELong-range evidenceDirectly evaluated
- SCState changesDirectly evaluated
- CFConflicting memoriesNot evaluated
- UQUnanswerable queriesNot evaluated
- MAMemory-grounded actionDirectly evaluated
- SMSpatial memoryDirectly evaluated
- OSObject-state trackingDirectly evaluated
- TSTask-state trackingDirectly evaluated
Survey notes
Multi-session interaction tests memory formation, maintenance, retrieval and use under changing world states.