Michael Rabinovich
readme: describe the space card in client-facing terms
6d31809
---
title: CADGenBench Leaderboard
emoji: 🔧
colorFrom: indigo
colorTo: pink
sdk: docker
app_port: 7860
pinned: true
license: apache-2.0
hf_oauth: true
short_description: Leaderboard for AI-driven CAD generation
datasets:
- HuggingAI4Engineering/cadgenbench-submissions
- HuggingAI4Engineering/cadgenbench-data
tags:
- leaderboard
- modality:3d
- submission:automatic
- test:private
- language:english
- eval:generation
- eval:cad
- judge:auto
---
# CADGenBench Leaderboard
A benchmark for AI-driven CAD generation: how well can a system turn a
textual or visual description of a mechanical part into a valid,
geometrically correct 3D model?
Browse the **Leaderboard** tab for ranked submissions, or upload your
own STEP files on the **Submit** tab: the Space runs the eval and
publishes a row with a per-submission report. The **About** tab covers
the scoring axes and links to the benchmark code and datasets.