Rule-Compliant Visual Spatial Planning for Multimodal Large Language Models

Yu Chen, Ting Lei, Yaoyi Li, Jia Cai, Zhecen Wu et al.|August 20, 2026arXiv

Key Takeaway

Breaking down rule-following tasks into separate perception, rule interpretation, and planning steps helps multimodal models generalize better to new rules and constraints than end-to-end approaches.

Summary

This paper introduces RuleMaze, a benchmark for testing whether multimodal AI models can navigate mazes while following natural-language rules. The authors propose a method called Disentangled Multimodal Planning that breaks down the task into separate steps—understanding the scene, interpreting rules, and planning actions—to help models follow complex constraints more reliably.

reasoning multimodal evaluation

Key Terms

multimodal-input spatial-reasoning constrained-decoding interpretable-reasoning