MMBench-Video
A long-form multi-shot benchmark for holistic video understanding that incorporates approximately 600 web videos from YouTube spanning 16 major categories, with each video ranging from 30 seconds to 6 minutes. Includes roughly 2,000 original question-answer pairs covering 26 fine-grained capabilities.
Qwen2.5 VL 72B Instruct from Alibaba Cloud / Qwen Team currently leads the MMBench-Video leaderboard with a score of 0.020 across 3 evaluated AI models.
What MMBench-Video measures
MMBench-Video is a multimodal benchmark that evaluates large language models on multimodal, reasoning, video, and vision tasks. LLM Stats tracks 3 models on this benchmark, with a maximum possible score of 1. Current average across reported models is 0.0, with the leader reaching 0.0.
Compare leaders on the best AI for multimodal, best AI for reasoning, best AI for video and best AI for vision leaderboards.
Publication
- Paper
- MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding
- Authors
- Xinyu Fang, Kangrui Mao, Haodong Duan, Xiangyu Zhao, and 3 others
- Published
- arXiv
- 2406.14515
Abstract
The advent of large vision-language models (LVLMs) has spurred research into their applications in multi-modal contexts, particularly in video understanding. Traditional VideoQA benchmarks, despite providing quantitative metrics, often fail to encompass the full spectrum of video content and inadequately assess models' temporal comprehension. To address these limitations, we introduce MMBench-Video, a quantitative benchmark designed to rigorously evaluate LVLMs' proficiency in video understanding. MMBench-Video incorporates lengthy videos from YouTube and employs free-form questions, mirroring practical use cases. The benchmark is meticulously crafted to probe the models' temporal reasoning skills, with all questions human-annotated according to a carefully constructed ability taxonomy. We employ GPT-4 for automated assessment, demonstrating superior accuracy and robustness over earlier LLM-based evaluations. Utilizing MMBench-Video, we have conducted comprehensive evaluations that include both proprietary and open-source LVLMs for images and videos. MMBench-Video stands as a valuable resource for the research community, facilitating improved evaluation of LVLMs and catalyzing progress in the field of video understanding. The evalutation code of MMBench-Video will be integrated into VLMEvalKit: https://github.com/open-compass/VLMEvalKit.
Qwen2.5 VL 72B Instruct leads with 2.0%, followed by Qwen2.5 VL 32B Instruct at 1.9% and Qwen2.5 VL 7B Instruct at 1.8%.
Progress Over Time
Interactive timeline showing model performance evolution on MMBench-Video
MMBench-Video Leaderboard
| Context | Cost | License | ||||
|---|---|---|---|---|---|---|
| 1 |  Alibaba Cloud / Qwen Team | 72B | — | — | ||
| 2 | Alibaba Cloud / Qwen Team | 34B | — | — | ||
| 3 | Alibaba Cloud / Qwen Team | 8B | — | — |
FAQ
Common questions about MMBench-Video.
More evaluations to explore
Related benchmarks in the same category
A challenging dataset of 448 multiple-choice questions written by domain experts in biology, physics, and chemistry. Questions are Google-proof and extremely difficult, with PhD experts reaching 65% accuracy.
A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks. Features over 12,000 curated questions across 14 domains and causes a 16-33% accuracy drop compared to original MMLU.
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
Massive Multitask Language Understanding benchmark testing knowledge across 57 diverse subjects including STEM, humanities, social sciences, and professional domains
A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.
Humanity's Last Exam (HLE) is a multi-modal academic benchmark with 2,500 questions across mathematics, humanities, and natural sciences, designed to test LLM capabilities at the frontier of human knowledge with unambiguous, verifiable solutions