Multimodal ai models tested on physics based bridge design and repair
PolyBridgeBench: Benchmarking Multimodal LLMs for Physics-Grounded Bridge Design
Artificial Intelligence
Summary
Many AI models can understand images and create designs, but it's unclear if their designs actually hold up in the real world. The authors developed PolyBridgeBench, a new test that shows if AI can design bridges that stand up to physics and fix them when they fail. The test uses visual and engineering rules to check if the bridge designs work, then simulates forces to see if they hold. The authors found that many AI models can make valid plans but often fail under real physics and struggle to fix mistakes with limited attempts.
What this means in practice
- •For structural engineers: Evaluate AI-generated bridge designs for real-world physics feasibility and automated repair suggestions before physical prototyping.
- •For civil engineering software developers: Incorporate multimodal AI benchmarks to improve tools that generate and validate structural designs under physics constraints.
Authors
Zicheng Zhao, Dongyin Chen, Rui Xu, Yinghui Xu
Abstract
Multimodal large language models, or MLLMs, perform well at visual understanding and structured generation, yet these capabilities do not establish whether an engineering design will work when executed. Existing benchmarks assess spatial reasoning, structural validity, or physics-grounded construction, but they do not determine whether MLLMs can synthesize complete load-bearing structures and repair them after simulator execution exposes a failure. We introduce PolyBridgeBench, an executable benchmark for multimodal bridge design. A model receives a visual scene and structured engineering constraints and generates a complete node--member--material topology. Deterministic legality checks gate execution in a native dynamic physics simulation. Following an execution failure, the benchmark returns temporal visual evidence from the failed rollout and evaluates repair under a fixed interaction budget. Separate measurements of deterministic validity, dynamic functional success, and post-failure recovery identify the stage at which design fails. Experiments with six representative MLLMs across 189 levels expose a substantial gap between deterministic validity and dynamic success, pronounced sensitivity to material budgets, and limited post-failure recovery under the primary strict-budget setting.