3DHarnessBench: Probing Agentic 3D-to-Code Capabilities of Frontier Vision-Language Models

2026-09-06

AI Summary

该研究通过3DHarnessBench评估前沿视觉-语言模型从多种输入生成Blender Python代码恢复3D几何的能力，采用逐步增强的探索设置。

Abstract

We introduce 3DHarnessBench, a benchmark that evaluates the agentic ability of frontier vision-language models (VLMs) to recover 3D geometry as Blender Python code from a variety of inputs. Unlike previous frameworks that prompt the VLMs with a fixed input (e.g., a single rendering or a text description), 3DHarnessBench evaluates four separate harness settings that progressively enable active agentic exploration, facilitated by recent Blender MCP functionality. Our hierarchy from Single-view, Multi-view, Active Visual (arbitrary viewpoint access), and Full 3D Interaction (complete access to the target object through Blender function calls) probes the models'abilities in both visual perception and active inference, tool calling, and self-correction. We observe that the ability of all frontier models to recover 3D geometry improves significantly with richer function call access, although the improvements are strongly model-dependent, revealing highly uneven agentic 3D-to-code capabilities. We will release the benchmark, code, outputs, and agent trajectories for reproducible 3D evaluation.

Problem

• 3D Geometry
• Vision-Language Models
• Agentic Exploration
• Blender Python Code
• Benchmark

Innovation

• 3D-to-Code
• Agentic Exploration
• Blender MCP
• Active Inference
• Self-Correction