A unified benchmark of 145 tasks across object, scene, and interactive open-world 3D creation evaluates coding agents on appearance, structure, and behavior, revealing reference fidelity and coordinated activity completion as central bottlenecks.
Coding agents combine visual understanding with program execution, offering a new route to automated 3D creation. By translating images and instructions into editable programs, they bring reconstruction, generation, and editing into a shared workflow. Yet existing benchmarks fragment evaluation across task scopes and protocols, leaving their capabilities across construction levels unclear. We propose Frontier3DBench, a unified benchmark of 145 tasks across 9 categories spanning object, scene, and interactive open-world creation, with joint evaluation of appearance, structure, and behavior. Experiments with 11 model-harness configurations reveal strengths that vary by level: GPT-6.1-sol…
coding agentsvisual understandingprogram executionautomated 3D creationeditable programsreconstructiongenerationeditingFrontier3DBenchobject creation
- Upvotes
- 6
- Authors
- 13
- Published
- Nov 6, 2026
Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.