UniEditBench: A Unified and Cost-Effective Benchmark for Image and Video Editing via Distilled MLLMs

• URL: http://arxiv.org/abs/2604.15871v1
• Date: Fri, 17 Apr 2026 09:21:48 GMT
• Title: UniEditBench: A Unified and Cost-Effective Benchmark for Image and Video Editing via Distilled MLLMs
• Abstract summary: We present UniEditBench, a unified benchmark for image and video editing.<n>UniEditBench supports reconstruction-based and instruction-driven methods under a shared protocol.<n>We distill a high-capacity MLLM judge into lightweight 4B/8B evaluators that provide multi-dimensional scoring over structural fidelity, text alignment, background consistency, naturalness, and temporal-spatial consistency for videos.
• Score: 19.634771714749544
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The evaluation of visual editing models remains fragmented across methods and modalities. Existing benchmarks are often tailored to specific paradigms, making fair cross-paradigm comparisons difficult, while video editing lacks reliable evaluation benchmarks. Furthermore, common automatic metrics often misalign with human preference, yet directly deploying large multimodal models (MLLMs) as evaluators incurs prohibitive computational and financial costs. We present UniEditBench, a unified benchmark for image and video editing that supports reconstruction-based and instruction-driven methods under a shared protocol. UniEditBench includes a structured taxonomy of nine image operations (Add, Remove, Replace, Change, Stroke-based, Extract, Adjust, Count, Reorder) and eight video operations, with coverage of challenging compositional tasks such as counting and spatial reordering. To enable scalable evaluation, we distill a high-capacity MLLM judge (Qwen3-VL-235B-A22B Instruct) into lightweight 4B/8B evaluators that provide multi-dimensional scoring over structural fidelity, text alignment, background consistency, naturalness, and temporal-spatial consistency (for videos). Experiments show that the distilled evaluators maintain strong agreement with human judgments and substantially reduce deployment cost relative to the teacher model. UniEditBench provides a practical and reproducible protocol for benchmarking modern visual editing methods. Our benchmark and the associated reward models are publicly available at https://github.com/wesar1/UniEditBench.

Related papers

• OmniEdit-Bench: A Comprehensive Benchmark for Instruction-based Video Editing [53.21868575804758]
  We introduce a comprehensive and structured benchmark for instruction-based video editing (IVE)<n>Our benchmark decomposes editing tasks into multiple video-specific dimensions, including spatial, temporal, audio, and reference-based editing.<n>We propose an evaluation framework that assesses editing quality from four complementary dimensions: accuracy, preservation, realism, and consistency.
  arXiv  Detail & Related papers  (2026-08-05T16:58:48Z)
• CoVEBench: Can Video Editing Models Handle Complex Instructions? [26.871333176496126]
  CoVEBench is a compositional video editing benchmark comprising 416 curated source videos, 626 multi-point editing instructions, and 9,990 fine-grained checklist items.<n>It evaluates models via MLLM-judged instruction compliance and video fidelity, alongside automated metrics for video quality.
  arXiv  Detail & Related papers  (2026-06-07T02:29:41Z)
• VEFX-Bench: A Holistic Benchmark for Generic Video Editing and Visual Effects [33.5974992507736]
  We introduce VEFX-Dataset, a human-annotated dataset containing 5,049 video editing examples across 9 major editing categories and 32 subcategories.<n> VEFX-Reward is a reward model designed specifically for video editing quality assessment.<n> VEFX-Bench is a benchmark of 300 curated video-prompt pairs for standardized comparison of editing systems.
  arXiv  Detail & Related papers  (2026-04-17T17:28:24Z)
• GEditBench v2: A Human-Aligned Benchmark for General Image Editing [58.86807672117726]
  We introduce GEditBench v2, a comprehensive benchmark with 1,200 real-world user queries spanning 23 tasks.<n>We also propose PVC-Judge, an open-source pairwise assessment model for visual consistency, trained via two novel region-decoupled preference data synthesis pipelines.<n> PVC-Judge achieves state-of-the-art evaluation performance among open-source models and even surpasses GPT-5.1 on average.
  arXiv  Detail & Related papers  (2026-03-30T15:08:32Z)
• DLEBench: Evaluating Small-scale Object Editing Ability for Instruction-based Image Editing Model [10.609050605838805]
  This paper introduces DeepLookEditBench, the first benchmark dedicated to assessing the abilities of IIEMs in editing small-scale objects.<n>We construct a challenging testbed comprising 1889 samples across seven instruction types.<n>In these samples, target objects occupy only 1%-10% of the image area, covering complex scenarios such as partial occlusion and multi-object editing.<n> Empirical results on 10 IIEMs reveal significant performance gaps in small-scale object editing, highlighting the need for specialized benchmarks to advance this ability.
  arXiv  Detail & Related papers  (2026-02-27T02:59:34Z)
• I2I-Bench: A Comprehensive Benchmark Suite for Image-to-Image Editing Models [78.62380562116135]
  Existing image editing benchmarks suffer from limited task scopes, insufficient evaluation dimensions, and heavy reliance on manual annotations.<n>We propose textbfI2I-Bench, a comprehensive benchmark for image-to-image editing models, which features 10 task categories across both single-image and multi-image editing tasks.<n>Using I2I-Bench, we benchmark numerous mainstream image editing models, investigating the gaps and trade-offs between editing models across various dimensions.
  arXiv  Detail & Related papers  (2025-12-04T10:44:07Z)
• UniREditBench: A Unified Reasoning-based Image Editing Benchmark [52.54256348710893]
  This work proposes UniREditBench, a unified benchmark for reasoning-based image editing evaluation.<n>It comprises 2,700 meticulously curated samples, covering both real- and game-world scenarios across 8 primary dimensions and 18 sub-dimensions.<n>We fine-tune Bagel on this dataset and develop UniREdit-Bagel, demonstrating substantial improvements in both in-domain and out-of-distribution settings.
  arXiv  Detail & Related papers  (2025-11-03T07:24:57Z)
• EdiVal-Agent: An Object-Centric Framework for Automated, Fine-Grained Evaluation of Multi-Turn Editing [170.71134330650796]
  EdiVal-Agent is an object-centric evaluation framework for instruction-based image editing.<n>It is designed to assess not only standard single-turn but also multi-turn instruction-based editing with precision.<n>We build EdiVal-Bench, a benchmark covering 9 instruction types and 13 state-of-the-art editing models spanning in-context, flow-matching, and diffusion paradigms.
  arXiv  Detail & Related papers  (2025-09-16T17:45:39Z)
• RefEdit: A Benchmark and Method for Improving Instruction-based Image Editing Model on Referring Expressions [56.9437856499838]
  We introduce RefEdit -- an instruction-based editing model trained on our scalable synthetic data generation pipeline.<n>Our RefEdit, trained on only 20,000 editing triplets, outperforms the Flux/SD3 model-based baselines trained on millions of data.
  arXiv  Detail & Related papers  (2025-06-03T23:20:24Z)
• CompBench: Benchmarking Complex Instruction-guided Image Editing [63.347846732450364]
  CompBench is a large-scale benchmark for complex instruction-guided image editing.<n>We propose an MLLM-human collaborative framework with tailored task pipelines.<n>We propose an instruction decoupling strategy that disentangles editing intents into four key dimensions.
  arXiv  Detail & Related papers  (2025-05-18T02:30:52Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.