# WorldAuditBench: Interactive 3D World Auditing with Multimodal Agents

arXiv: https://arxiv.org/abs/2609.40325 · HTML: https://mlsanity.com/paper/2609.40325/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 22 of its 31 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)
- LLM agents and tool use / [Multimodal and embodied agents](https://mlsanity.com/agents/multimodal-and-embodied-agents/)

## Builds on (most cited first, 22 of 22)

- [Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis](https://mlsanity.com/paper/2405.21075/) (2024)
- [Habitat: A Platform for Embodied AI Research](https://mlsanity.com/paper/1904.01201/) (2019)
- [AI2-THOR: An Interactive 3D Environment for Visual AI](https://mlsanity.com/paper/1712.05474/) (2017)
- [Holodeck: Language Guided Generation of 3D Embodied AI Environments](https://mlsanity.com/paper/2312.09067/) (2023)
- [EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents](https://mlsanity.com/paper/2502.09560/) (2025)
- [PhyScene: Physically Interactable 3D Scene Synthesis for Embodied AI](https://mlsanity.com/paper/2404.09465/) (2024)
- [VisualAgentBench: Towards Large Multimodal Models as Visual Foundation Agents](https://mlsanity.com/paper/2408.06327/) (2024)
- [3D-Mem: 3D Scene Memory for Embodied Exploration and Reasoning](https://mlsanity.com/paper/2411.17735/) (2024)
- [UnrealCV: Connecting Computer Vision to Unreal Engine](https://mlsanity.com/paper/1609.01326/) (2016)
- [Vid2Sim: Realistic and Interactive Simulation from Video for Urban Navigation](https://mlsanity.com/paper/2501.06693/) (2025)
- [UnrealZoo: Enriching Photo-realistic Virtual Worlds for Embodied AI](https://mlsanity.com/paper/2412.20977/) (2024)
- [GlitchBench: Can large multimodal models detect video game glitches?](https://mlsanity.com/paper/2312.05291/) (2023)
- [SimWorld: An Open-ended Realistic Simulator for Autonomous Agents in Physical and Social Worlds](https://mlsanity.com/paper/2512.01078/) (2025)
- [VideoGameBunny: Towards vision assistants for video games](https://mlsanity.com/paper/2407.15295/) (2024)
- [LAYOUTDREAMER: Physics-guided Layout for Text-to-3D Compositional Scene Generation](https://mlsanity.com/paper/2502.01949/) (2025)
- [VideoGameQA-Bench: Evaluating Vision-Language Models for Video Game Quality Assurance](https://mlsanity.com/paper/2505.15952/) (2025)
- [Scaling Behavior Cloning Improves Causal Reasoning: An Open Model for Real-Time Video Game Playing](https://mlsanity.com/paper/2601.04575/) (2026)
- [ClutterGen: A Cluttered Scene Generator for Robot Learning](https://mlsanity.com/paper/2407.05425/) (2024)
- [FlySearch: Exploring how vision-language models explore](https://mlsanity.com/paper/2506.02896/) (2025)
- [We'll Fix It in Post: What Do Bug Fixes in Video Game Update Notes Tell Us?](https://mlsanity.com/paper/2103.03997/) (2021)
- [Open-Ended Video Game Glitch Detection with Agentic Reasoning and Temporal Grounding](https://mlsanity.com/paper/2604.07818/) (2026)
- [Deriving and Evaluating a Detailed Taxonomy of Game Bugs](https://mlsanity.com/paper/2311.16645/) (2023)

## Abstract (arXiv)

As interactive 3D worlds are increasingly used to study intelligent behavior, it becomes important to develop efficient pipelines for identifying anomalies in these simulated environments, such as floating objects, traversable walls, or objects inconsistent with the surrounding scene. Multimodal AI systems, including vision-language models (VLMs) and vision-language-action models (VLAs), have shown potential for automating this task. However, 3D world auditing is complex, requiring the close coupling of two distinct capabilities: action, to navigate the 3D world and search for anomalies systematically and efficiently; and visual reasoning, to understand the environment and identify anomalies from multimodal observations. It remains largely unexplored whether multimodal agents can effectively couple these two capabilities, using visual reasoning to identify potential anomalies while taking actions to validate them. In this paper, we introduce WorldAuditBench, a benchmark for 3D world auditing comprising 213 anomaly tasks across 13 environments built with Unreal Engine 5 and Three.js, spanning five anomaly families. We evaluate five frontier models under a fixed exploration budget using two auditing paradigms: VLA-based exploration followed by VLM-based anomaly identification, and an end-to-end VLM agent in which visual reasoning directly guides action selection. Across the evaluated models and two paradigms, success rates range from 6.6% to 42.3%, substantially below human performance (83.4%). Through the task of world auditing, WorldAuditBench provides a testbed for studying how multimodal agents couple action and visual reasoning in interactive 3D environments, while highlighting current limitations in their ability to gather and interpret evidence during exploration.