Artifact-Bench: Evaluating MLLMs on Detecting and Assessing the Artifacts of AI-Generated Videos
Quick Take
Artifact-Bench evaluates MLLMs on detecting artifacts in AI-generated videos, revealing significant limitations.
Key Points
- Introduces a benchmark for artifact detection in videos.
- Establishes a taxonomy of realism artifacts.
- Finds MLLMs struggle with artifact perception and reasoning.
📖 Reader Mode
~2 min readAuthors:Yuqi Tang, Yang Shi, Zhuoran Zhang, Qixun Wang, Xuehai Bai, Yue Ding, Ruizhe Chen, Bohan Zeng, Xinlong Chen, Xuanyu Zhu, Bozhou Li, Yuran Wang, Yifan Dai, Chengzhuo Tong, Xinyu Liu, Yiyan Ji, Yujie Wei, Yuhao Dong, Shilin Yan, Fengxiang Wang, Yi-Fan Zhang, Haotian Wang, Yuanxing Zhang, Pengfei Wan
Abstract:Recent video generative models have greatly improved the realism of AI-generated videos, yet their outputs still exhibit artifacts such as temporal inconsistencies, structural distortions, and semantic incoherence. While Multimodal Large Language Models (MLLMs) show strong visual understanding capabilities, their ability to perceive and reason about such artifacts remains unclear. Existing benchmarks often lack systematic evaluation of artifact-aware perception and fine-grained diagnostic reasoning, especially across diverse AI-generated video domains beyond photorealistic content. To address this gap, we introduce Artifact-Bench, a comprehensive benchmark for evaluating MLLMs on AI-generated video artifact detection and analysis. We first establish a three-level hierarchical taxonomy of realism artifacts, covering photorealistic, animated, and CG-style videos. Based on this taxonomy, Artifact-Bench defines three complementary tasks: real vs. AI-generated video classification, pairwise realism comparison, and fine-grained artifact identification. Experiments on 19 leading MLLMs reveal substantial limitations in artifact perception and reasoning, with many models approaching random or even below-random performance in challenging settings. We further observe significant misalignment between MLLM judgments and human perceptual preferences, highlighting their limited reliability as general evaluators for AI-generated video realism.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV) |
| Cite as: | arXiv:2605.18984 [cs.CV] |
| (or arXiv:2605.18984v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2605.18984 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yang Shi [view email]
[v1]
Mon, 18 May 2026 18:04:54 UTC (21,044 KB)
— Originally published at arxiv.org
More from arXiv cs.CV
See more →GeoSym127K: Scalable Symbolically-verifiable Synthesis for Multimodal Geometric Reasoning
GeoSym127K introduces a scalable neuro-symbolic framework for enhanced geometric reasoning in multimodal models.