VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender
cs.CV, cs.AI
Submitted: 2026-11-11
Updated: 2026-11-18
Project page: https://vibeprevisbench.github.io
Terminology
Sources
- VideoPhy-2: A Challenging Action-Centric Physical Commonsense Evaluation in Video Generation
- Lighting-grounded Video Generation with Renderer-based Agent Reasoning
- PVBench: Quantifying Physical Plausibility in Video Generation with Object-Level Physical Measurements
- World Narrative Model for Highly Controllable Video Generation: A Paradigm Shift from Pixel Sampling to Physical World Orchestration
- Code World Model: Coding Agent as World Brain
- When Left Is Right? Camera-Motion Understanding beyond Screen
- VibeVideoBench: Evaluating Multimodal Agents on Code-Driven Video Generation
- PACE: Precise AI Cinematic Expression
- BlenderWorld: Benchmarking Visual Agents for 3D Creation from Real-World References
- HumanScore: Benchmarking Human Motions in Generated Videos
- Code Plans, Diffusion Renders: Open-Ended Generative World Modeling
- 3DCodeBench: Benchmarking Agentic Procedural 3D Modeling Via Code
- CaMo: Camera Motion Grounded Evaluation and Training for Vision-Language Models
- PROWBench: Do Video Models Render What the Program Specifies?
- StoryBlender: Inter-Shot Consistent and Editable 3D Storyboard with Spatial-temporal Dynamics
- WorldModelBench: Judging Video Generation Models As World Models
- VideoCoCo: Code-as-CoT for Physically-Consistent Video Generation via an Agentic Dual-Engine System
- CinemaTraj: Composing Atomic Camera Trajectories for 3D Scenes with LLM Agents
- 3D Primitives are a Spatial Language for VLMs
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs