Research map / Large language models / Agents, code and math
Op-CAD: Benchmarking and Investigating Operation-oriented CAD Generation
Yixue Bai, Yufei Gu, Zeke Xie
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add geometry guidance to annotation pipeline (ablation) | annotation pipeline · Annotation quality (1000 samples) | GPT-4o preference: 31.0% -> 50.5% | Improved |
| add step-level annotation (shape) stage (ablation) | annotation pipeline · Annotation quality (1000 samples) | GPT-4o preference: 18.5% -> 50.5% | Improved |
| add geometry guidance to annotation pipeline (ablation) | annotation pipeline · Annotation quality (1000 samples) | human preference: 25.0% -> 62.5% | Improved |
| add step-level annotation (shape) stage (ablation) | annotation pipeline · Annotation quality (1000 samples) | human preference: 12.5% -> 62.5% | Improved |
| fine-tune Llama-3.1-8B on Op-CAD (Op-Llama-8B) | Llama-3.1-8B · Op-CAD SSM | CF-IoU: 0.044 -> 0.120 | Improved |
| fine-tune Llama-3.1-8B on Op-CAD (Op-Llama-8B) | Llama-3.1-8B · Op-CAD SSM | Pass@1: 0.624 -> 0.783 | Improved |
| fine-tune Llama-3.1-8B on Op-CAD (Op-Llama-8B) | Llama-3.1-8B · Op-CAD SGM | Pass@2: 0.849 -> 0.835 | Hurt |
| fine-tune Qwen3-8B on Op-CAD | Qwen3-8B · Op-CAD SSM | Pass@1: 0.639 -> 0.742 | Improved |
| fine-tune Qwen3-8B on Op-CAD | Qwen3-8B · Op-CAD SSM | Mean CD: 20.549 -> 24.348 | Hurt |
| add COOP prompting | Llama-3.1-8B · Op-CAD SSM | Pass@1: 0.624 -> 0.645 | Improved |
| add COOP prompting | GPT-4o · Op-CAD SSM | CF-IoU: 0.133 -> 0.108 | Hurt |
| add COOP prompting | Op-Llama-8B · Op-CAD SSM | Pass@1: 0.783 -> 0.817 | Improved |
Closest papers by meaning
- CAD-Editor: A Locate-then-Infill Framework with Automated Training Data Synthesis for Text-Based CAD Editing (ICML 2025 · similarity 0.88)
- CAD-Llama: Leveraging Large Language Models for Computer-Aided Design Parametric 3D Model Generation (CVPR 2025 · similarity 0.87)
- GeoCAD: Local Geometry-Controllable CAD Generation with Large Language Models (NeurIPS 2025 · similarity 0.87)
- CAD-GPT: Synthesising CAD Construction Sequence with Spatial Reasoning-Enhanced Multimodal LLMs (AAAI 2025 · similarity 0.87)
- CAD-Coder: Text-to-CAD Generation with Chain-of-Thought and Geometric Reward (NeurIPS 2025 · similarity 0.87)
- Generating CAD Code with Vision-Language Models for 3D Designs (ICLR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
