MIND: Benchmarking Memory Consistency and Action Control in World Models

TL;DR

MIND is a comprehensive benchmark designed to evaluate memory consistency and action control in world models through diverse videos and an interactive baseline, highlighting current challenges in long-term memory and action generalization.

Contribution

This paper introduces MIND, the first unified benchmark for assessing core abilities of world models in dynamic environments, including a new evaluation framework and an interactive baseline.

Findings

Current world models struggle with long-term memory consistency.

Generalization across different action spaces remains challenging.

MIND provides a comprehensive platform for future research and benchmarking.

Abstract

World models aim to understand, remember, and predict dynamic visual environments, yet a unified benchmark for evaluating their fundamental abilities remains lacking. To address this gap, we introduce MIND, the first open-domain closed-loop revisited benchmark for evaluating Memory consIstency and action coNtrol in worlD models. MIND contains 250 high-quality videos at 1080p and 24 FPS, including 100 (first-person) + 100 (third-person) video clips under a shared action space and 25 + 25 clips across varied action spaces covering eight diverse scenes. We design an efficient evaluation framework to measure two core abilities: memory consistency and action control, capturing temporal stability and contextual coherence across viewpoints. Furthermore, we design various action spaces, including different character movement speeds and camera rotation angles, to evaluate the action generalization capability across different action spaces under shared scenes. To facilitate future performance benchmarking on MIND, we introduce MIND-World, a novel interactive Video-to-World baseline. Extensive experiments demonstrate the completeness of MIND and reveal key challenges in current world models, including the difficulty of maintaining long-term memory consistency and generalizing across action spaces. Code: https://github.com/CSU-JPG/MIND.