The DOI matches its Crossref record; 54 of the 54 references it cites were found in a registry or an index.
MLVU: Benchmarking Multi-task Long Video Understanding
Open accessAbstract
The evaluation of Long Video Understanding (LVU) performance poses an important but challenging research problem. Despite previous efforts, the existing video understanding benchmarks are severely constrained by several issues, especially the insufficient lengths of videos, a lack of diversity in video types and evaluation tasks, and the inappropriateness for evaluating LVU performances. To address the above problems, we propose a new benchmark called MLVU (Multitask Long Video Understanding Benchmark) for the comprehensive and in-depth evaluation of LVU. MLVU presents the following critical values: 1) The substantial and flexible extension of video lengths, which enables the benchmark to evaluate LVU performance across a wide range of durations. 2) The inclusion of various video genres, such as movies, surveillance, egocentric videos, and cartoons, reflects the models’ LVU performances in different scenarios. 3) The development of diversified evaluation tasks, which enables a comprehensive examination of MLLMs’ key abilities in long-video understanding. The empirical study with 23 latest MLLMs reveals significant room for improvement in today’s technique, as all existing methods struggle with most of the evaluation tasks and exhibit severe performance degradation when handling longer videos. Additionally, it suggests that factors such as context length, image-understanding ability, and the choice of LLM backbone can play critical roles in future advancements. We anticipate that MLVU will advance the research of LVU by providing a comprehensive and in-depth analysis of MLLMs. The code and dataset can be accessed from https://github.com/JUNJIE99/MLVU.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Most existing LVQA benchmarks [2, 11, 50, 62] mainly focus on visual understanding, such as action recognition and spatial/temporal perception, leaving a gap in evaluating higher-level cognitive abilities that rely on long-term memory and are crucial for real-world agents, such as understanding persons, extracting general knowledge, and performing cross-modal reasoning.”
From Seeing, Listening, Remembering, and Reasoning: A Multimodal Agent with Long-Term Memory (2025) · cites this paper as [62]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Research works printed but not linked: 8
- [6][6] Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Language models are few-shot learners. Advances in neural information processing systems, 33:1877-1901, 2020. 1 -
- [15][15] Kristen Grauman, Andrew Westbury, Eugene Byrne, Zachary Chavis, Antonino Furnari, Rohit Girdhar, Jackson Hamburger, Hao Jiang, Miao Liu, Xingyu Liu, et al. Ego4d: Around the world in 3,000 hours of egocentric video. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 18995-19012, 2022. 1, 3, 4 -
- [34][34] Karttikeya Mangalam, Raiymbek Akshulakov, and Jitendra Malik. Egoschema: A diagnostic benchmark for very longform video language understanding. Advances in Neural Information Processing Systems, 36, 2023. 1, 2, 3 -
- [48][48] Bo Wu, Shoubin Yu, Zhenfang Chen, Joshua B Tenenbaum, and Chuang Gan. Star: A benchmark for situated reasoning in real-world videos. In Thirty-fifth conference on neural information processing systems datasets and benchmarks track (Round 2), 2021. 3 -
- [52][52] Jun Xu, Tao Mei, Ting Yao, and Yong Rui. Msr-vtt: A large video description dataset for bridging video and language. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5288-5296, 2016. 1, 2, 3 -
- [55][55] Kexin Yi, Chuang Gan, Yunzhu Li, Pushmeet Kohli, Jiajun Wu, Antonio Torralba, and Joshua B Tenenbaum. Clevrer: Collision events for video representation and reasoning. In International Conference on Learning Representations, 2019. 4 -
- [56][56] Zhou Yu, Dejing Xu, Jun Yu, Ting Yu, Zhou Zhao, Yueting Zhuang, and Dacheng Tao. Activitynet-qa: A dataset for understanding complex web videos via question answering. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 9127-9134, 2019. 3 -
- [63][63] Deyao Zhu, Jun Chen, Xiaoqian Shen, Xiang Li, and Mohamed Elhoseiny. Minigpt-4: Enhancing visionlanguage understanding with advanced large language models. In The Twelfth International Conference on Learning Representations, 2023. 2
Cited by
Sorted by citations
Items 101–200 of 454
Exploring the Effect of Reinforcement Learning on Video Understanding: Insights from SEED-Bench-R1Preprint
Chen, Y., Ge, Y., Wang, R., Ge, Y., Qiu, L., Shan, Y., & Liu, X. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.24376 · Cited by 25
EgoExoBench: A Benchmark for First- and Third-person View Video Understanding in MLLMs
He, Y., Huang, Y., Chen, G., Pei, B., Xu, J., Lü, T., & Pang, J. (2025) · neural information processing systems · doi:10.52202/085713-3891 · Cited by 24
LinVT: Empower Your Image-level Large Language Model to Understand VideosPreprint
Gao, L., Zhong, Y., Zeng, Y., Tan, H., Li, D., & Zhao, Z. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2412.05185 · Cited by 24
Memory-enhanced Retrieval Augmentation for Long Video UnderstandingPreprint
Yuan, H., Liu, Z., Qin, M., Qian, H., Yan, S., Dou, Z., Wen, J.-R., & Sebe, N. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.09149 · Cited by 24
Self-citationQuality assessment in the era of large models: A survey
Zhang, Z., Zhou, Y., Li, C., Zhao, B., Liu, X., & Zhai, G. (2025) · ACM Transactions on Multimedia Computing, Communications, and Applications · doi:10.1145/3722559 · Cited by 23
Self-citationSTAR-R1: Spatial TrAnsformation Reasoning by Reinforcing Multimodal LLMsPreprint
Li, Z., Ma, Z., Li, M., Li, S., Rong, Y., Xu, T., Zhang, Z., Zhao, D., & Huang, W. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.15804 · Cited by 23
Active Video Perception: Iterative Evidence Seeking for Agentic Long Video UnderstandingPreprint
Wang, Z., Zhou, H., Wang, S., Li, J., Xiong, C., Savarese, S., Bansal, M., Ryoo, M. S., & Niebles, J. C. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.05774 · Cited by 23
TSPO: Temporal Sampling Policy Optimization for Long-form Video Language Understanding
Tang, C., Han, Z., Sun, H., Zhou, S., Zhang, X., Wei, X., Yuan, Y., Zhang, H., Xu, J., & Sun, H. (2026) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v40i11.37896 · Cited by 22
VRoPE: Rotary Position Embedding for Video Large Language Models
Liu, Z., Guo, L., Tang, Y., Yue, T., Cai, J., Ma, K., Liu, Q., Chen, X., & Liu, J. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-main.731 · Cited by 22
Self-citationLong Story Short: Story-level Video Understanding from 20K Short Films
Ghermi, R., Wang, X., Kalogeiton, V. S., & Laptev, I. (2026) · International Journal of Computer Vision · doi:10.1007/s11263-026-02940-x · Cited by 22
SF2T: Self-supervised Fragment Finetuning of Video-LLMs for Fine-Grained UnderstandingPreprint
Hu, Y., Song, Z., Feng, N., Luo, Y., Yu, J., Chen, Y. P., & Yang, W. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.07745 · Cited by 22
BIMBA: Selective-Scan Compression for Long-Range Video Question Answering
Islam, M. M., Nagarajan, T., Wang, H., Bertasius, G., & Torresani, L. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02709 · Cited by 21
Wang, Z., Yoon, J., Yu, S., Islam, M. M., Bertasius, G., & Bansal, M. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-main.1428 · Cited by 21
Does Your Vision-Language Model Get Lost in the Long Video Sampling Dilemma?
Qu, T., Tang, L., Peng, B., Yang, S., Yu, B., & Jia, J. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01942 · Cited by 21
VideoZoomer: Reinforcement-Learned Temporal Focusing for Long Video ReasoningPreprint
Ding, Y., Zhang, Y., Lai, X., Chu, R., & Yang, Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.22315 · Cited by 21
MemLens: Benchmarking Multimodal Long-Term Memory in Large Vision-Language ModelsPreprint
Ren, X., Wang, Z., Du, Y., Xie, Z., Liu, C., Yang, X., Feng, H., Pan, W., Zheng, T., Xu, B., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.14906 · Cited by 22
Self-citationAeroDuo: Aerial Duo for UAV-based Vision and Language Navigation
Wu, R., Zhang, Y., Chen, J., Huang, L., Zhang, S., Xu, Z. Y., Wang, L., & Liu, S. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3754498 · Cited by 20
Multi-Granular Spatio-Temporal Token Merging for Training-Free Acceleration of Video LLMs
Hyun, J., Hwang, S., Han, S. H., Kim, T., Lee, I., Wee, D., Lee, J., Kim, S. J., & Shim, M. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02224 · Cited by 20
VRBench: A Benchmark for Multi-Step Reasoning in Long Narrative Videos
Yu, J., Wu, Y., Chu, M., Ren, Z., Huang, Z., Chu, P., Zhang, R., He, Y., Li, Q., Li, S., et al. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02011 · Cited by 20
MolmoPoint: Better Pointing for VLMs with Grounding Tokens
Clark, C., Yang, Y., Park, J. S., Ma, Z., Zhang, J., Tripathi, R., Salehi, M., Lee, S., Anderson, T., Han, W., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37242-0_37 · Cited by 20
MME-VideoOCR: Evaluating OCR-Based Capabilities of Multimodal LLMs in Video Scenarios
Shi, Y., Wang, H., Xie, W., Zhang, H., Zhao, L., Zhang, Y., Li, X., Fu, C., Wen, Z., Liu, W., et al. (2025) · neural information processing systems · doi:10.52202/085713-2779 · Cited by 20
VideoEval-Pro: Robust and Realistic Long Video Understanding EvaluationPreprint
Ma, W., Ren, W., Jia, Y., Li, Z., Nie, P., Zhang, G., & Chen, W. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.14640 · Cited by 20
DenseWorld-1M: Towards Detailed Dense Grounded Caption in the Real WorldPreprint
Li, X., Zhang, T., Li, Y., Yuan, H., Chen, S., Zhou, Y., Meng, J., Sun, Y., Xu, S., Qi, L., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2506.24102 · Cited by 20
Ye, J., Gong, S., Gao, J., Fan, J., Wu, S., Bi, W., Bai, H., Shang, L., & Kong, L. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.22615 · Cited by 20
Zhang, Y., Liu, X., Tao, R., Chen, Q., Fei, H., Che, W., & Qin, L. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3755837 · Cited by 19
Thus Spake Long-Context Large Language ModelPreprint
Liu, X., Li, R., Huang, M., Liu, Z., Song, Y., Guo, Q., He, S., Wang, Q., Li, L., Liu, Q., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2502.17129 · Cited by 19
Self-citationBreaking Down Video LLM Benchmarks: Knowledge, Spatial Perception, or True Temporal Understanding?Preprint
Feng, B., Lai, Z., Li, S., Wang, Z., Wang, S., Huang, P., & Cao, M. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.14321 · Cited by 19
X-LeBench: A Benchmark for Extremely Long Egocentric Video Understanding
Zhou, W., Cao, K., Zheng, H., Liu, Y., Zheng, X., Liu, M., Kristensson, P. O., Mayol-Cuevas, W. W., Zhang, F., Lin, W., et al. (2025) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2025.findings-emnlp.822 · Cited by 18
SITE: Towards Spatial Intelligence Thorough Evaluation
Wang, W., Tan, R., Zhu, P., Yang, J., Yang, Z., Wang, L., Kolobov, A., Gao, J., & Gong, B. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.00847 · Cited by 18
Unveiling the Ignorance of MLLMs: Seeing Clearly, Answering Incorrectly
Liu, Y., Liang, Z., Wang, Y., Wu, X., Tang, F., He, M., Li, J., Liu, Z., Yang, H., Lim, S.-N., et al. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.00849 · Cited by 18
Self-citationTask Preference Optimization: Improving Multimodal Large Language Models with Vision Task Alignment
Yan, Z., Li, Z., He, Y., Wang, C., Li, K., Li, X., Zeng, X., Wang, Z., Wang, Y., Qiao, Y., et al. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02781 · Cited by 18
T*: Re-thinking Temporal Search for Long-Form Video UnderstandingPreprint
Ye, J., Wang, Z., Sun, H., Chandrasegaran, K., Durante, Z., Eyzaguirre, C., Bisk, Y., Niebles, J. C., Adeli, E., Fei-Fei, L., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.02259 · Cited by 18
PACT: Pruning and Clustering-Based Token Reduction for Faster Visual Language Models
Dhouib, M., Buscaldi, D., Haddad-Vanier, S., & Shabou, A. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.01359 · Cited by 17
FlexSelect: Flexible Token Selection for Efficient Long Video Understanding
zhang, Y., Lu, Y. Y., Wang, T., Rao, F., Yang, Y., & Zhu, L. (2025) · neural information processing systems · doi:10.52202/085713-3435 · Cited by 17
From Frames to Clips: Training-free Adaptive Key Clip Selection for Long-Form Video UnderstandingPreprint
Sun, G., Singhal, A., Uzkent, B., Shah, M., Chen, C., & Kessler, G. N. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.02262 · Cited by 17
MMR-V: What's Left Unsaid? A Benchmark for Multimodal Deep Reasoning in VideosPreprint
Zhu, K., Jin, Z., Yuan, H., Li, J., Tu, S., Cao, P., Chen, Y., Liu, K., & Zhao, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2506.04141 · Cited by 17
ToDRE: Effective Visual Token Pruning via Token Diversity and Task Relevance
Li, D., Yang, Z., Zhang, X., Shao, L., & Lu, S. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37167-6_36 · Cited by 16
Ren, W., Yang, H., Min, J., Wei, C., & Chen, W. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.00360 · Cited by 16
STEP: Enhancing Video-LLMs’ Compositional Reasoning by Spatio-Temporal Graph-guided Self-Training
Qiu, H., Gao, M., Qian, L., Pan, K., Yu, Q., Li, J., Wang, W., Tang, S., Zhuang, Y., & Chua, T. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.00312 · Cited by 16
StreamKV: Streaming Video Question-Answering with Segment-based KV Cache Retrieval and CompressionPreprint
Chen, Y., Bai, X., Wang, Z., Bai, C., Dai, Y., Lu, M. Z., & Zhang, S. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2511.07278 · Cited by 20
Pan, J., Zhang, Q., Zhang, R., Lu, M., Wan, X., Zhang, Y., Liu, C., & She, Q. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2511.05489 · Cited by 16
Yuan, L., Wang, J., Sun, H., Zhang, Y., & Lin, Y. (2025) · Qeios · doi:10.32388/x26ilu · Cited by 15
ReFoCUS: Reinforcement-guided Frame Optimization for Contextual UnderstandingPreprint
Lee, H., Kim, J., Kim, H., & Ro, Y. M. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2506.01274 · Cited by 15
QuoTA: Query-oriented Token Assignment via CoT Query Decouple for Long Video ComprehensionPreprint
Luo, Y., Chen, W., Zheng, X., Huang, W., Yin, S., Lin, H., Fu, C., Huang, J., Ji, J., Luo, J., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.08689 · Cited by 15
VideoScan: Enabling Efficient Streaming Video Understanding via Frame-level Semantic CarriersPreprint
Li, R., Tan, Y., Shi, Y., & Shao, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.09387 · Cited by 15
Mage-VL: An Efficient Codec-Native Streaming Multimodal Foundation ModelPreprint
Yang, S., Zhang, K., Jia, Z., Guo, J., Shen, Y., Zhang, X., Zhang, X., Wang, H., Li, X., Zhang, P., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2607.24904 · Cited by 15
Self-citationMulti-Agent System for Comprehensive Soccer Understanding
Rao, J., Li, Z., Wu, H., Zhang, Y., Wang, Y., & Xie, W. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3755144 · Cited by 14
HoPE: Hybrid of Position Embedding for Long Context Vision-Language Models
Li, H., Qin, Y., Ou, B., Xu, L., & Xu, R. (2025) · neural information processing systems · doi:10.52202/085713-5430 · Cited by 14
ST-VLM: Kinematic Instruction Tuning for Spatio-Temporal Reasoning in Vision-Language ModelsPreprint
Ko, D., Kim, S., Suh, Y., G, V. K. B., Yoon, M., Chandraker, M. K., & Kim, H. J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.19355 · Cited by 14
MMG-Vid: Maximizing Marginal Gains at Segment-level and Token-level for Efficient Video LLMsPreprint
Ma, J., Zhang, Q., Lu, M. F., Wang, Z., Zhou, Q., Song, J. L., & Zhang, S. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2508.21044 · Cited by 14
Mixed-R1: Unified Reward Perspective For Reasoning Capability in Multimodal Large Language ModelsPreprint
Xu, S., Li, Y., Yang, R., Zhang, T., Sun, Y., Chow, W., Li, L., Song, H., Xu, Q., Tong, Y., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.24164 · Cited by 14
CrossLMM: Decoupling Long Video Sequences from LMMs via Dual Cross-Attention MechanismsPreprint
Yan, S., Han, J., Tsai, J., Xue, H., Fang, R., Hong, L., Guo, Z., & Zhang, R. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.17020 · Cited by 14
What Limits Vision-and-Language Navigation ?Preprint
Wang, Y., Fang, Y., Wang, T., Li, L., Liu, K., Xu, J., Yuan, Z., Feng, Y., Zhang, J., Lu, W., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.13328 · Cited by 14
MOSS-VL Technical ReportPreprint
Wang, P., Tan, C., Zhou, S., Zhou, Q., Chen, Y., He, X., Zeng, H., Cheng, J., Wang, C., Qian, X., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2608.15045 · Cited by 14
Yan, S., Lin, H., Liu, Y., Zhang, Y., Zeng, G., Li, Y., Zhou, Y., Lim, S.-N., Yang, H. T., & Sebe, N. (2025) · neural information processing systems · doi:10.52202/085713-2320 · Cited by 13
MomentSeeker: A Task-Oriented Benchmark For Long-Video Moment Retrieval
Yuan, H., Ni, J., Liu, Z., Wang, Y., Zhou, J., Liang, Z., Zhao, B., Cao, Z., Wen, J.-R., & Dou, Z. (2025) · neural information processing systems · doi:10.52202/085713-0934 · Cited by 13
Self-citationInternVideo3: Agentify Foundation Models with Multimodal Contextual ReasoningPreprint
Yan, Z., Xia, S., Yu, J., Wu, Y., Jiang, T., Li, S., Tian, K., Xu, Y., He, Y., (129938), K. C., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2606.12195 · Cited by 13
Qu, T., Li, M., Tuytelaars, T., & Moens, M. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2411.11066 · Cited by 12
VerIPO: Cultivating Long Reasoning in Video-LLMs via Verifier-Gudied Iterative Policy OptimizationPreprint
Li, Y., Chen, X., Li, Z., Liu, Z., Wang, L., Luo, W., Hu, B., & Min, Z. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.19000 · Cited by 12
Self-citationDynTok: Dynamic Compression of Visual Tokens for Efficient and Effective Video UnderstandingPreprint
Zhang, H., Zhang, J., Ji, X., Wang, Q., & Zhang, F. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2506.03990 · Cited by 12
Zhao, Y., Zeng, Y., Qi, Y., Liu, Y., Bao, X., Chen, L., Chen, Z., Miao, Q., Liu, C., Zhao, J., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.17736 · Cited by 12
OneVision-Encoder: Codec-Aligned Sparsity as a Foundational Principle for Multimodal IntelligencePreprint
Tang, F., An, X., Yan, Y., Xie, Y., Qin, B., Yang, K., Shen, Y., Zhang, Y., Li, C., Feng, S., et al. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2602.08683 · Cited by 12
Self-citationSmolVLM: Redefining small and efficient multimodal models
Marafioti, A., Zohar, O., Farré, M., Noyan, M., Bakouch, E., Cuenca, P., Zakka, C., Allal, L. B., Lozhkov, A., Tazi, N., et al. (2025) · arXiv · doi:10.48550/arxiv.2504.05299 · Cited by 11
See&Trek: Training-Free Spatial Prompting for Multimodal Large Language Model
Li, P., Song, P., Li, W., Yao, H., Guo, W., Xu, Y., Liu, D., & Xiong, H. (2025) · neural information processing systems · doi:10.52202/085713-2076 · Cited by 11
The ethical evaluation of large language models and its optimization
Lyu, Y., & Du, Y. (2025) · AI and Ethics · doi:10.1007/s43681-024-00654-9 · Cited by 11
Question-Answering Dense Video Events
Qin, H., Xiao, J., & Yao, A. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3729945 · Cited by 11
LongViTU: Instruction Tuning for Long-Form Video UnderstandingPreprint
Wu, R., Ma, X., Ci, H., Fan, Y., Wang, Y., Zhao, H., Li, Q., & Wang, Y. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2501.05037 · Cited by 11
ViaRL: Adaptive Temporal Grounding via Visual Iterated Amplification Reinforcement LearningPreprint
Xu, Z., Dai, Q., Xie, T., Yang, Y., Qiu, K., Chen, D., Wu, Z., & Luo, C. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.15447 · Cited by 11
QuickVideo: Real-Time Long Video Understanding with System Algorithm Co-DesignPreprint
Schneider, B., Jiang, D., Du, C., Pang, T., & Chen, W. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.16175 · Cited by 11
VidText: Towards Comprehensive Evaluation for Video Text UnderstandingPreprint
Yang, Z., Shu, Y., Wang, J., Yang, Z., Zhang, Y., Li, Y., Lu, K., Zeng, G., Liu, S., Zhou, Y., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.22810 · Cited by 11
Self-citationLess Is More, but Where? Dynamic Token Compression via LLM-Guided Keyframe PriorPreprint
Li, Y., Gui, H., Fan, Z., Wang, J., Kang, B., Chen, B., & Tian, Z. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.06866 · Cited by 11
Vid-SME: Membership Inference Attacks against Large Video Understanding Models
Li, Q., Yu, R., & Wang, X. (2025) · neural information processing systems · doi:10.52202/085713-3725 · Cited by 10
CausalStep: A Benchmark for Explicit Stepwise Causal Reasoning in Videos
Li, X., Li, X., Hu, S., Huang, K., & Zhang, W. (2026) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v40i8.37582 · Cited by 10
AdaVideoRAG: Omni-Contextual Adaptive Retrieval-Augmented Efficient Long Video Understanding
Xue, Z., Zhang, J., Xurong, X., Cai, Y., Liu, Y., Li, X., & Tao, D. (2025) · neural information processing systems · doi:10.52202/085713-0213 · Cited by 10
iLLaVA: An Image is Worth Fewer Than 1/3 Input Tokens in Large Multimodal ModelsPreprint
Hu, L., Gao, L., Shang, F., Wan, L., & Feng, W. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2412.06263 · Cited by 10
LLaVA-Octopus: Unlocking Instruction-Driven Adaptive Projector Fusion for Video UnderstandingPreprint
Sun, B., Zhao, J., Chen, X., Wei, X., & Hou, Q. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2501.05067 · Cited by 10
An LMM for Efficient Video Understanding via Reinforced Compression of Video CubesPreprint
Qi, J., Yao, Y., Bai, Y., Xu, B., Li, J., Liu, Z., & Chua, T.-S. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.15270 · Cited by 10
Self-citationSlow-Fast Architecture for Video Multi-Modal Large Language ModelsPreprint
Shi, M., Wang, S., Chen, C.-Y., Jain, J., Wang, K., Xiong, J., Liu, G., Yu, Z., & Shi, H. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.01328 · Cited by 10
IPCV: Information-Preserving Compression for MLLM Visual EncodersPreprint
Chen, Y., Wen, Z., Wu, Y., Liu, X., Chen, S., Ma, J., Li, W., He, C., & Zhang, L. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.18747 · Cited by 10
StreamReady: Learning What to Answer and When in Long Streaming VideosPreprint
Azad, S., Vineet, V., & Rawat, Y. S. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.08620 · Cited by 10
METok: Multi-Stage Event-based Token Compression for Efficient Long Video Understanding
Wang, M., Chen, S., Kersting, K., Volker Tresp, C., & Ma, Y. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-main.954 · Cited by 9
Cross-Modal and Uncertainty-Aware Agglomeration for Open-Vocabulary 3D Scene Understanding
Li, J., Saltori, C., Poiesi, F., & Sebe, N. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.01806 · Cited by 9
CROP: Contextual Region-Oriented Visual Token Pruning
Guo, J., Zhai, F., Pu, J., Wei, Q., & Zhou, Y. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-main.492 · Cited by 10
Suo, Y., Ma, F., Zhu, L., Wang, T., Rao, F., & Yang, Y. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.02158 · Cited by 9
LightVLM: Acceleraing Large Multimodal Models with Pyramid Token Merging and KV Cache CompressionPreprint
Hu, L., Shang, F., Feng, W., & Wan, L. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.00419 · Cited by 9
Event-Anchored Frame Selection for Effective Long-Video UnderstandingPreprint
Chen, W., Luo, Y., Zeng, Y., Lin, L., Xie, T., Chao, F., Ji, R., & Zheng, X. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.00983 · Cited by 9
Wavelet-based Frame Selection by Detecting Semantic Boundary for Long Video UnderstandingPreprint
Chen, W., Zeng, Y., Luo, Y., Xie, T., Lin, L., Ji, J., Zhang, Y., & Zheng, X. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.00512 · Cited by 9
V-CAST: Video Curvature-Aware Spatio-Temporal Pruning for Efficient Video Large Language ModelsPreprint
Lin, X., Liu, X., Wang, Y., Ma, T., & Ren, W. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.27650 · Cited by 9
ALLVB: All-in-One Long Video Understanding Benchmark
Tan, X., Luo, Y., Ye, Y., Liu, F., & Cai, Z. (2025) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v39i7.32775 · Cited by 8
Towards Video Thinking Test: A Holistic Benchmark for Advanced Video Reasoning and Understanding
Zhang, Y., Chew, Y., Dong, Y., Leo, A., Hu, B., & Liu, Z. (2025) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51701.2025.01918 · Cited by 8
Self-citationZoomV: Temporal Zoom-in for Efficient Long Video Understanding
Zhang, Y., Pan, J., Zhang, R., Wan, X., Zhang, Q., Lu, M., She, Q., & Zhang, S. (2026) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3767308.3835256 · Cited by 8
Small Vision-Language Models are Smart Compressors for Long Video Understanding
Fei, J., Chen, J., Liu, Z., Xiong, Y., Zhou, C., Wen, W., Han, J., Zhuge, M., Suri, S., Qian, Q., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37189-8_8 · Cited by 8
Self-citationReflect-R1: Evidence-Driven Reflection for Self-correction in Long Video Understanding
Chen, S., Chen, Y., Guan, Y., Cheng, Z., Zhang, Z., Qin, S., Xia, B., Li, J., Yang, W., & Ma, F. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37464-6_25 · Cited by 8
Chowdhury, S., Elmoghany, M., Abeysinghe, Y., Fei, J., Nag, S., Khan, S. A., Elhoseiny, M., & Manocha, D. (2025) · neural information processing systems · doi:10.52202/085713-1648 · Cited by 8
An Evolved Universal Transformer MemoryPreprint
Cetin, E., Sun, Q., Zhao, T., & Tang, Y. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2410.13166 · Cited by 8
RAG-Adapter: A Plug-and-Play RAG-enhanced Framework for Long Video UnderstandingPreprint
Tan, X., Ye, Y., Luo, Y., Wan, Q., Liu, F., & Cai, Z. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2503.08576 · Cited by 8
ScaleLong: A Multi-Timescale Benchmark for Long Video UnderstandingPreprint
Ma, D., Yuan, H., Wang, X., Zang, Q., Liu, T., He, X., Wei, Y., Guo, J., Jiahui, N., Yang, Z., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.23922 · Cited by 8
Threading Keyframe with Narratives: MLLMs as Strong Long Video ComprehendersPreprint
Fang, B., Wu, W., Wu, Q., Song, Y., & Chan, A. B. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.24158 · Cited by 8
AuroraLong: Bringing RNNs Back to Efficient Open-Ended Video UnderstandingPreprint
Xu, W., Song, E., Chai, W., Wen, X., Tian, Y., & Wang, G. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2507.02591 · Cited by 8
ActionArt: Advancing Multimodal Large Models for Fine-Grained Human-Centric Video UnderstandingPreprint
Peng, Y.-X., Yang, Q., Tang, Y.-M., Fu, S., Lin, K., Wei, X., & Zheng, W. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2504.18152 · Cited by 8
Metadata
- Authors
- Zhou, J., Shu, Y., Zhao, B., Wu, B., Liang, Z., Xiao, S., Qin, M., Yang, X., Xiong, Y., Zhang, B., Huang, T., & Liu, Z.
- Affiliations
- Beijing University of Posts and Telecommunications (CN)
- Beijing Academy of Artificial Intelligence (CN)
- Shanghai Jiao Tong University (CN)
- Zhejiang University (CN)
- ORCID
- Venue
- 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2025-06-10
- Pages
- 13691-13701
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W4413144845
- Cited by
- 44
- References
- 0
- Citations by year
- 2026: 37 · 2025: 6
- FWCI
- 32.37
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion, cc-by-nc-sa)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Human Pose and Action Recognition
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Video Analysis and Summarization
- Keywords
- long video understanding, benchmark, multi-task evaluation, multimodal large language models, video length, video genres, surveillance video, egocentric video, animated cartoons, context length, image understanding
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
