aiqbench analysis
This research introduces MecEng, an automated benchmark assessing Large Language Models' ability to generate multibody simulation models from text descriptions involving complex spatial geometry and mechanics. The study evaluates 34 LLMs across rigid-body and flexible systems tasks, revealing that while performance on simpler problems is high, accurately modeling flexible components remains a significant challenge for current architectures. These findings provide critical insight into the specific limitations of AI in engineering domains where precise physical reasoning is required rather than just pattern matching.
Large Language Models and their Awareness of Mechanics and Spatial Geometry
arXiv:2608.14615v1 Announce Type: new Abstract: Large Language Models (LLMs) perform well on established code-generation and mathematical-reasoning benchmarks, but their capabilities in mechanics and spatial geometry, here denoted as mechanical engineering awareness, has not been quantified systematically. We present MecEng, a fully automated benchmark that evaluates LLMs on the creation of multibody simulation models from parameterized textual descriptions. The benchmark comprises 84 generic tasks on three difficulty levels, ranging from rigid-body systems with joints and contact to flexible multibody systems that require exact 3D geometry generation, tetrahedral finite-element meshing, and Hurty-Craig-Bampton model order reduction of machine parts. A dedicated pipeline with LLMs generates simulation-ready geometry from text using Netgen, and builds multibody system models for the code Exudyn, which are then verified against expert ground truth on several levels: system-graph isomorphism including graph node annotations, numerical solutions, and part-specific measures such as mass, geometry, and eigenfrequencies. In total, 32 open-weight and two proprietary LLMs are evaluated. On rigid-body tasks, the best open-weight model obtains an overall success rate of 86.0%, compared to 91.4% for the strongest proprietary model, while flexible multibody tasks remain considerably harder. Additional studies quantify the influence of sampling temperature, reasoning, prompt design, model size, and LLM-release date. The results indicate rapidly improving, but still error-prone, mechanical engineering awareness of current LLMs.
Abstract:Large Language Models (LLMs) perform well on established code-generation and mathematical-reasoning benchmarks, but their capabilities in mechanics and spatial geometry, here denoted as mechanical engineering awareness, has not been quantified systematically. We present MecEng, a fully automated benchmark that evaluates LLMs on the creation of multibody simulation models from parameterized textual descriptions. The benchmark comprises 84 generic tasks on three difficulty levels, ranging from rigid-body systems with joints and contact to flexible multibody systems that require exact 3D geometry generation, tetrahedral finite-element meshing, and Hurty-Craig-Bampton model order reduction of machine parts. A dedicated pipeline with LLMs generates simulation-ready geometry from text using Netgen, and builds multibody system models for the code Exudyn, which are then verified against expert ground truth on several levels: system-graph isomorphism including graph node annotations, numerical solutions, and part-specific measures such as mass, geometry, and eigenfrequencies. In total, 32 open-weight and two proprietary LLMs are evaluated. On rigid-body tasks, the best open-weight model obtains an overall success rate of 86.0%, compared to 91.4% for the strongest proprietary model, while flexible multibody tasks remain considerably harder. Additional studies quantify the influence of sampling temperature, reasoning, prompt design, model size, and LLM-release date. The results indicate rapidly improving, but still error-prone, mechanical engineering awareness of current LLMs.
Artificial Intelligence (cs.AI)
arXiv:2608.14615 [cs.AI]
(or arXiv:2608.14615v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2608.14615 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 301 words
Related coverage
Robot-GST: geometry-aware spatial-temporal robot policy representation and evaluation
arXiv · Sep 30, 2026
J++ Lens: Jacobian Filtering Enables More Faithful Workspace Lenses
LessWrong · Oct 8, 2026
The Machines that Make the Machines
NVIDIA · Oct 7, 2026
StrataVLA: Hierarchical and Efficient 3D Geometric Grounding for Vision-Language-Action Models
arXiv · Sep 24, 2026