Text2CAD-Bench: A Benchmark for LLM-based Text-to-Parametric CAD Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917508155441152 |
|---|---|
| author | Wang, Liang Meng, Heng Xiang, Zekai Liu, Jin Zhou, Pingyi Chen, Litao Tang, Yongqiang |
| author_facet | Wang, Liang Meng, Heng Xiang, Zekai Liu, Jin Zhou, Pingyi Chen, Litao Tang, Yongqiang |
| contents | Text-to-CAD generation aims to create parametric CAD models from natural language, enabling rapid prototyping and intuitive design workflows. However, existing benchmarks focus on basic primitives and simple sketch-extrude sequences, lacking advanced features essential for real-world applications and covering only traditional mechanical parts. We introduce Text2CAD-Bench, the first benchmark systematically evaluating text-to-CAD across geometric complexity and application diversity. Our benchmark comprises 600 human-curated examples spanning four levels: L1-L2 cover fundamental geometry with standard features, L3 introduces complex topology and freeform surfaces, and L4 extends to real-world domains beyond mechanical parts. Each example pairs dual-style prompts -- geometric descriptions mimicking non-expert users, and procedural sequences aligned with expert-level conventions. Evaluating mainstream general LLMs and domain-specific models, we find that current models perform reasonably on basic geometry but degrade substantially on complex topology and advanced features. We release our benchmark to drive progress in text-to-CAD research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_18430 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Text2CAD-Bench: A Benchmark for LLM-based Text-to-Parametric CAD Generation Wang, Liang Meng, Heng Xiang, Zekai Liu, Jin Zhou, Pingyi Chen, Litao Tang, Yongqiang Machine Learning Text-to-CAD generation aims to create parametric CAD models from natural language, enabling rapid prototyping and intuitive design workflows. However, existing benchmarks focus on basic primitives and simple sketch-extrude sequences, lacking advanced features essential for real-world applications and covering only traditional mechanical parts. We introduce Text2CAD-Bench, the first benchmark systematically evaluating text-to-CAD across geometric complexity and application diversity. Our benchmark comprises 600 human-curated examples spanning four levels: L1-L2 cover fundamental geometry with standard features, L3 introduces complex topology and freeform surfaces, and L4 extends to real-world domains beyond mechanical parts. Each example pairs dual-style prompts -- geometric descriptions mimicking non-expert users, and procedural sequences aligned with expert-level conventions. Evaluating mainstream general LLMs and domain-specific models, we find that current models perform reasonably on basic geometry but degrade substantially on complex topology and advanced features. We release our benchmark to drive progress in text-to-CAD research. |
| title | Text2CAD-Bench: A Benchmark for LLM-based Text-to-Parametric CAD Generation |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2605.18430 |