CADEvolve: Creating Realistic CAD via Program Evolution

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Elistratov, Maksim, Barannikov, Marina, Ivanov, Gregory, Khrulkov, Valentin, Konushin, Anton, Kuznetsov, Andrey, Zhemchuzhnikov, Dmitrii
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908839551434752
author Elistratov, Maksim
Barannikov, Marina
Ivanov, Gregory
Khrulkov, Valentin
Konushin, Anton
Kuznetsov, Andrey
Zhemchuzhnikov, Dmitrii
author_facet Elistratov, Maksim
Barannikov, Marina
Ivanov, Gregory
Khrulkov, Valentin
Konushin, Anton
Kuznetsov, Andrey
Zhemchuzhnikov, Dmitrii
contents Computer-Aided Design (CAD) delivers rapid, editable modeling for engineering and manufacturing. Recent AI progress now makes full automation feasible for various CAD tasks. However, progress is bottlenecked by data: public corpora mostly contain sketch-extrude sequences, lack complex operations, multi-operation composition and design intent, and thus hinder effective fine-tuning. Attempts to bypass this with frozen VLMs often yield simple or invalid programs due to limited 3D grounding in current foundation models. We present CADEvolve, an evolution-based pipeline and dataset that starts from simple primitives and, via VLM-guided edits and validations, incrementally grows CAD programs toward industrial-grade complexity. The result is 8k complex parts expressed as executable CadQuery parametric generators. After multi-stage post-processing and augmentation, we obtain a unified dataset of 1.3m scripts paired with rendered geometry and exercising the full CadQuery operation set. A VLM fine-tuned on CADEvolve achieves state-of-the-art results on the Image2CAD task across the DeepCAD, Fusion 360, and MCB benchmarks.
format Preprint
id arxiv_https___arxiv_org_abs_2602_16317
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle CADEvolve: Creating Realistic CAD via Program Evolution
Elistratov, Maksim
Barannikov, Marina
Ivanov, Gregory
Khrulkov, Valentin
Konushin, Anton
Kuznetsov, Andrey
Zhemchuzhnikov, Dmitrii
Graphics
Computer-Aided Design (CAD) delivers rapid, editable modeling for engineering and manufacturing. Recent AI progress now makes full automation feasible for various CAD tasks. However, progress is bottlenecked by data: public corpora mostly contain sketch-extrude sequences, lack complex operations, multi-operation composition and design intent, and thus hinder effective fine-tuning. Attempts to bypass this with frozen VLMs often yield simple or invalid programs due to limited 3D grounding in current foundation models. We present CADEvolve, an evolution-based pipeline and dataset that starts from simple primitives and, via VLM-guided edits and validations, incrementally grows CAD programs toward industrial-grade complexity. The result is 8k complex parts expressed as executable CadQuery parametric generators. After multi-stage post-processing and augmentation, we obtain a unified dataset of 1.3m scripts paired with rendered geometry and exercising the full CadQuery operation set. A VLM fine-tuned on CADEvolve achieves state-of-the-art results on the Image2CAD task across the DeepCAD, Fusion 360, and MCB benchmarks.
title CADEvolve: Creating Realistic CAD via Program Evolution
topic Graphics
url https://arxiv.org/abs/2602.16317