# CADBench: A Multimodal Benchmark for AI-Assisted CAD Program Generation

arXiv: https://arxiv.org/abs/2605.10873 · HTML: https://mlsanity.com/paper/2605.10873/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 28 of its 34 references that are in the map.
- First posted on arXiv 11 May 2026.

## Builds on (most cited first, 28 of 28)

- [Objaverse: A Universe of Annotated 3D Objects](https://mlsanity.com/paper/2212.08051/) (2022)
- [DINOv3](https://mlsanity.com/paper/2508.10104/) (2025)
- [ABC: A Big CAD Model Dataset For Geometric Deep Learning](https://mlsanity.com/paper/1812.06216/) (2018)
- [DeepCAD: A Deep Generative Network for Computer-Aided Design Models](https://mlsanity.com/paper/2105.09492/) (2021)
- [BRepNet: A topological message passing system for solid models](https://mlsanity.com/paper/2104.00706/) (2021)
- [CAD-MLLM: Unifying Multimodality-Conditioned CAD Generation With MLLM](https://mlsanity.com/paper/2411.04954/) (2024)
- [SketchGen: Generating Constrained CAD Sketches](https://mlsanity.com/paper/2106.02711/) (2021)
- [CAD-Recode: Reverse Engineering CAD Code from Point Clouds](https://mlsanity.com/paper/2412.14042/) (2024)
- [CAD-SIGNet: CAD Language Inference from Point Clouds using Layer-wise Sketch Instance Guided Attention](https://mlsanity.com/paper/2402.17678/) (2024)
- [Generating CAD Code with Vision-Language Models for 3D Designs](https://mlsanity.com/paper/2410.05340/) (2024)
- [Point2Cyl: Reverse Engineering 3D Objects from Point Clouds to Extrusion Cylinders](https://mlsanity.com/paper/2112.09329/) (2021)
- [CAD-GPT: Synthesising CAD Construction Sequence with Spatial Reasoning-Enhanced Multimodal LLMs](https://mlsanity.com/paper/2412.19663/) (2024)
- [From Concept to Manufacturing: Evaluating Vision-Language Models for Engineering Design](https://mlsanity.com/paper/2311.12668/) (2023)
- [Text-to-CadQuery: A New Paradigm for CAD Generation with Scalable Large Model Capabilities](https://mlsanity.com/paper/2505.06507/) (2025)
- [CAD-Coder: An Open-Source Vision-Language Model for Computer-Aided Design Code Generation](https://mlsanity.com/paper/2505.14646/) (2025)
- [How Can Large Language Models Help Humans in Design and Manufacturing?](https://mlsanity.com/paper/2307.14377/) (2023)
- [GenCAD: Image-Conditioned Computer-Aided Design Generation with Transformer-Based Contrastive Representation and Diffusion Priors](https://mlsanity.com/paper/2409.16294/) (2024)
- [TransCAD: A Hierarchical Transformer for CAD Sequence Inference from Point Clouds](https://mlsanity.com/paper/2407.12702/) (2024)
- [BlenderLLM: Training Large Language Models for Computer-Aided Design with Self-improvement](https://mlsanity.com/paper/2412.14203/) (2024)
- [CADDreamer: CAD Object Generation from Single-view Images](https://mlsanity.com/paper/2502.20732/) (2025)
- [PvDeConv: Point-Voxel Deconvolution for Autoencoding CAD Construction in 3D](https://mlsanity.com/paper/2101.04493/) (2021)
- [Large Language Models for Computer-Aided Design: A Survey](https://mlsanity.com/paper/2505.08137/) (2025)
- [cadrille: Multi-modal CAD Reconstruction with Reinforcement Learning](https://mlsanity.com/paper/2505.22914/) (2025)
- [GenCAD-3D: CAD Program Generation using Multimodal Latent Space Alignment and Synthetic Dataset Balancing](https://mlsanity.com/paper/2509.15246/) (2025)
- [Zero-to-CAD: Agentic Synthesis of Interpretable CAD Programs at Million-Scale Without Real Data](https://mlsanity.com/paper/2604.24479/) (2026)
- [CADEvolve: Creating Realistic CAD via Program Evolution](https://mlsanity.com/paper/2602.16317/) (2026)
- [MiCADangelo: Fine-Grained Reconstruction of Constrained CAD Models from 3D Scans](https://mlsanity.com/paper/2510.23429/) (2025)
- [CADFit: Precise Mesh-to-CAD Program Generation with Hybrid Optimization](https://mlsanity.com/paper/2605.01171/) (2026)

## Built on by (most cited first, 11 of 11)

- [CADEngBench: It Looks Like CAD, but Does It Work? Evaluating Parametric Design, Assembly Reasoning, and Physics Simulation](https://mlsanity.com/paper/2608.09296/) (2026)
- [3DHarnessBench: Probing Agentic 3D-to-Code Capabilities of Frontier Vision-Language Models](https://mlsanity.com/paper/2609.06535/) (2026)
- [CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design](https://mlsanity.com/paper/2609.16251/) (2026)
- [CADForge: Agentic Single-View CAD Reconstruction with Explicit Geometry Reasoning](https://mlsanity.com/paper/2610.04262/) (2026)
- [RealCADBench: Benchmarking Parametric CAD Modeling from Industrial Design Intents](https://mlsanity.com/paper/2609.03773/) (2026)
- [VisCAD: A Foundation Model Suite with Multimodal Industrial CAD Intelligence](https://mlsanity.com/paper/2609.03811/) (2026)
- [MechReason: Benchmarking Multi-Image Multi-Hop Reasoning in Mechanical Engineering](https://mlsanity.com/paper/2609.16012/) (2026)
- [ConCAD: Constraint-Aware Image-to-CAD Generation with Dual-Granularity Rewards](https://mlsanity.com/paper/2609.34494/) (2026)
- [StepCAD: Mesh-to-CAD Code Generation via LLM Policy and Geometry-Guided Search](https://mlsanity.com/paper/2610.03799/) (2026)
- [CADFather: Autonomous CAD Reconstruction through Coordinated Tool Use](https://mlsanity.com/paper/2610.09127/) (2026)
- [PhysCodeGym: Benchmarking Coding Agents for Physically Verifiable 4D Reconstruction](https://mlsanity.com/paper/2610.17288/) (2026)

## Abstract (arXiv)

Recovering editable CAD programs from images or 3D observations is central to AI-assisted design, but progress is difficult to measure because existing evaluations are fragmented across datasets, modalities, and metrics. We introduce CADBench, a unified benchmark for multimodal CAD program generation. CADBench contains 18,000 evaluation samples spanning six benchmark families derived from DeepCAD, Fusion 360, ABC, MCB, and Objaverse; five input modalities including clean meshes, noisy meshes, single-view renders, photorealistic renders, and multi-view renders; and six metrics covering geometric fidelity, executability, and program compactness. STEP-based families are stratified by B-rep face count and all families are diversity-sampled to support controlled analysis across complexity and object variation. We benchmark eleven CAD-specialized and general-purpose vision-language systems, generating more than 1.4 million CAD programs. Under idealized inputs, specialized mesh-to-CAD models substantially outperform code-generating VLMs, which remain far from reliable CAD program reconstruction. CADBench further reveals three recurring failure modes: reconstruction quality degrades with geometric complexity, CAD-specialized models can be brittle under modality shift, and model rankings change across metrics. Together, these results position CADBench as a diagnostic testbed for measuring progress in editable 3D reconstruction and multimodal CAD understanding. The benchmark is publicly available at https://github.com/anniedoris/CADBench.