ExeWorldBench: Benchmarking Code-Based Executable 3D World Modeling
ExeWorldBench is a benchmark for executable 3D world modeling that evaluates how models translate physical understanding into executable, code-based Blender simulations. It defines three progressively harder tasks: Mechanism Executability (single mechanism from an image), Compositional Executability (multi-stage causal chains), and Intent-Driven Executable World Synthesis (component- or goal-constrained). Experiments show substantial limitations in executable world modeling.
