VideoRewardBench: Comprehensive Evaluation of Multimodal Reward Models for Video Understanding

• URL: http://arxiv.org/abs/2509.00484v1
• Date: Sat, 30 Aug 2025 12:50:55 GMT
• Title: VideoRewardBench: Comprehensive Evaluation of Multimodal Reward Models for Video Understanding
• Abstract summary: Multimodal reward models (MRMs) play a crucial role in the training, inference, and evaluation of Large Vision Language Models (LVLMs)<n>Existing benchmarks for evaluating MRMs in the video domain suffer from a limited number and diversity of questions.<n>We introduce VideoRewardBench, the first comprehensive benchmark covering four core aspects of video understanding.
• Score: 19.54215281137561
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multimodal reward models (MRMs) play a crucial role in the training, inference, and evaluation of Large Vision Language Models (LVLMs) by assessing response quality. However, existing benchmarks for evaluating MRMs in the video domain suffer from a limited number and diversity of questions, a lack of comprehensive evaluation dimensions, and inadequate evaluation of diverse types of MRMs. To address these gaps, we introduce VideoRewardBench, the first comprehensive benchmark covering four core aspects of video understanding: perception, knowledge, reasoning, and safety. Through our AI-assisted data pipeline, we curate a high-quality preference dataset of 1,563 annotated samples, including 1,482 unique videos and 1,559 distinct questions--15 times the number found in the most question-rich prior benchmark. Each sample is a triplet consisting of a video-text prompt, a chosen response, and a rejected response. We also conduct a comprehensive evaluation across 28 multimodal reward models spanning three categories: generative, discriminative, and semi-scalar. Results show that even the top-performing model GPT-4o achieves only 57.0% overall accuracy, and the state-of-the-art open-source model Qwen2.5-VL-72B reaches merely 53.3%. Our analysis further reveals three key insights: (i) MRMs trained with reinforcement learning (RL) do not necessarily exhibit stronger cross-modal generalization than those trained without RL; (ii) except for discriminative MRMs, other types of MRMs across varying model capacities can benefit from inference-time scaling; and (iii) variations in input video frame count have different effects on different types of MRMs. We believe VideoRewardBench offers a challenging and valuable benchmark for advancing the evaluation and development of MRMs in the video domain.

Related papers

• Modality Maturity Index: A benchmark for assessing multimodal capabilities of omni models [5.886031439164172]
  We propose the Modality Maturity Index (MMI) to evaluate the multimodal capabilities of large language models.<n>MMI consists of 893 questions, each carefully crafted to require the model to demonstrate its understanding of multiple input modalities.
  arXiv  Detail & Related papers  (2026-08-26T18:48:52Z)
• Video Understanding Reward Modeling: A Robust Benchmark and Performant Reward Models [51.616920646492524]
  We propose a unified framework spanning benchmark design, data construction, and reward model training.<n>We introduce Video Understanding Reward Bench (VURB), a benchmark featuring 2,100 preference pairs with long chain-of-thought reasoning traces (averaging 1,143 tokens) and majority voting evaluation across general, long, and reasoning-oriented video tasks.<n>We train VideoDRM and VideoGRM, a discriminative and a generative reward model, both achieving state-of-the-art performance on VURB and VideoRewardBench.
  arXiv  Detail & Related papers  (2026-05-08T15:29:11Z)
• MVU-Eval: Towards Multi-Video Understanding Evaluation for Multimodal LLMs [61.70050081221131]
  MVU-Eval is the first comprehensive benchmark for evaluating Multi-Video Understanding for MLLMs.<n>Our MVU-Eval mainly assesses eight core competencies through 1,824 meticulously curated question-answer pairs spanning 4,959 videos.<n>These capabilities are rigorously aligned with real-world applications such as multi-sensor synthesis in autonomous systems and cross-angle sports analytics.
  arXiv  Detail & Related papers  (2025-11-10T16:02:33Z)
• HV-MMBench: Benchmarking MLLMs for Human-Centric Video Understanding [79.06209664703258]
  Multimodal Large Language Models (MLLMs) have demonstrated significant advances in visual understanding tasks involving both images and videos.<n>Existing human-centric benchmarks predominantly emphasize video generation quality and action recognition, while overlooking essential perceptual and cognitive abilities required in human-centered scenarios.<n>We propose a rigorously curated benchmark designed to provide a more holistic evaluation of MLLMs in human-centric video understanding.
  arXiv  Detail & Related papers  (2025-07-07T11:52:24Z)
• VisualPRM: An Effective Process Reward Model for Multimodal Reasoning [76.35753243272521]
  We introduce VisualPRM, which improves the reasoning abilities of existing Multimodal Large Language Models (MLLMs)<n>Our model achieves a 5.9-point improvement across seven multimodal reasoning benchmarks.<n>For the evaluation of multimodal PRMs, we propose VisualProcessBench, a benchmark with human-annotated step-wise correctness labels.
  arXiv  Detail & Related papers  (2025-03-13T12:03:37Z)
• VLRMBench: A Comprehensive and Challenging Benchmark for Vision-Language Reward Models [40.87249469370042]
  Vision-language models (VLRMs) have become increasingly pivotal in the reasoning process.<n>Existing benchmarks for vision-language RMs (VLRMs) typically assess only a single aspect of their capabilities.<n>We propose a comprehensive and challenging benchmark, dubbed as VLRMBench, encompassing 12,634 questions.
  arXiv  Detail & Related papers  (2025-03-10T15:52:57Z)
• Multimodal RewardBench: Holistic Evaluation of Reward Models for Vision Language Models [82.92771279118888]
  We introduce Multimodal RewardBench, an expert-annotated benchmark for evaluating multimodal reward models.<n>Our dataset comprises 5,211 annotated (prompt, chosen response, rejected response) triplets collected from various vision-language models.<n>We find that even the top-performing models, Gemini 1.5 Pro and Claude 3.5 Sonnet, achieve only 72% overall accuracy.
  arXiv  Detail & Related papers  (2025-02-20T01:48:13Z)
• TOMATO: Assessing Visual Temporal Reasoning Capabilities in Multimodal Foundation Models [55.48403691519395]
  TOMATO is a novel benchmark crafted to rigorously assess MFMs' temporal reasoning capabilities in video understanding.<n>TOMATO comprises 1,484 carefully curated, human-annotated questions spanning six tasks.<n>Our comprehensive evaluation reveals a human-model performance gap of 57.3% with the best-performing model.
  arXiv  Detail & Related papers  (2024-10-30T17:50:23Z)
• Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis [120.67048724315619]
  Video-MME is the first-ever full-spectrum, Multi-Modal Evaluation benchmark of MLLMs in Video analysis.<n>We extensively evaluate various state-of-the-art MLLMs, including GPT-4 series and Gemini 1.5 Pro, as well as open-source image models.<n>Our experiments reveal that Gemini 1.5 Pro is the best-performing commercial model, significantly outperforming the open-source models.
  arXiv  Detail & Related papers  (2024-05-31T17:59:47Z)

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.