Text2CAD-Bench: A Benchmark for LLM-based Text-to-Parametric CAD Generation

2026-05-18

AI Summary

This work addresses the limitations of existing benchmarks for text-to-parametric CAD generation, which are confined to simple geometries and conventional parts and thus fail to evaluate performance on complex topologies, freeform surfaces, and multi-domain applications. We propose the first systematic evaluation benchmark comprising four difficulty levels—from basic geometric primitives to real-world complex designs—and introduce a novel dual-style prompting mechanism tailored for both non-expert and expert users. Leveraging 600 high-quality, manually curated examples, we assess mainstream general-purpose and domain-specific large language models, revealing that while they perform adequately on elementary tasks, their capabilities degrade significantly when handling intricate topologies and advanced modeling features. This study bridges the critical gap in evaluating parametric CAD generation along the dimensions of complexity and practical applicability.

Abstract

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.

Problem

• text-to-CAD
• parametric CAD
• benchmark
• geometric complexity
• application diversity

Innovation

• Text-to-CAD
• parametric CAD
• benchmark
• large language models
• geometric complexity