ACADEMICarXiv · 2025

CAD-GPT

CAD-GPT is a multimodal LLM fine-tuned to generate CAD construction sequences from both text and image inputs. Focuses on spatial reasoning — understanding 3D relationships from 2D projections.

Input
Text + image
Output
CAD sequences
Venue
arXiv
Year
2025
Links
Paper →Project page →
CAD Arena Status
Scheduled for evaluation on the full 200-prompt benchmark. Results will appear on the leaderboard at launch.