4DCodeBench: Benchmarking Agents on Inverse Graphics of Dynamic Scenes
cs.CV, cs.AI, cs.GR
Submitted: 2026-10-02
Updated: 2026-10-02
Comments: https://4dcodebench.com/
Code: https://github.com/4DCodeBench/4DCodeBench
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Scalable Behavior Cloning with Open Data, Training, and Evaluation
- IntPhys 2: Benchmarking Intuitive Physics Understanding In Complex Synthetic Environments
- 3DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Code
- Thinking in Blender: Staged Executable Inverse Graphics with Vision-Language Models
- GEOPHYS: The Geometry of Physical Plausibility
- MoGe-3: Fine-Detail Monocular Geometry Estimation with Self-Guided Sparse Volumetric Refinement
- MPMWorlds: Material-Point-Method Simulations for Inferring and Extrapolating Physical Dynamics
- VisPhyWorld: Probing Physical Reasoning via Code-Driven Video Reconstruction
- Robo360: A 3D Omnispective Multi-Material Robotic Manipulation Dataset
- SimWorlds: A Multi-Agent System for Dynamic 3D Scene Creation
- T-Rex: Tactile-Reactive Dexterous Manipulation
- DINOv3
- ManipArena: Comprehensive Real-world Evaluation of Reasoning-Oriented Generalist Robot Manipulation
- BVB: Benchmarking Agentic Video Understanding via Programmatic Reconstruction in Blender
- PhysCodeBench: Benchmarking Physics-Aware Symbolic Simulation of 3D Scenes via Self-Corrective Multi-Agent Refinement
- SceneActBench: Can Agents Act on the 3D Scenes They See?
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models