BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References

A new benchmark, BlenderWorld, evaluates multimodal large language model (MLLM) agents on creating and editing 3D models in Blender from real-world references, requiring the agents to infer object shapes, materials, spatial layouts and lighting from diverse reference inputs. The benchmark targets the gap between agents' ability to manipulate 3D software via code or direct control and the practical workflow demand of reconstructing scenes from real-world imagery.

Recent advances in multimodal large language models MLLMs have enabled them to act as autonomous agents, creating and editing 3D models via code or direct software manipulation. In practical workflows, these models must infer object shapes, materials, spatial layouts, and lighting from diverse ref