LongInsightBench: A Comprehensive Benchmark for Evaluating Omni-Modal Models on Human-Centric Long-Video Understanding.

We introduce LongInsightBench, the first benchmark designed to assess models' ability to understand long videos, with a focus on human language, viewpoints, actions, and other contextual elements, while integrating visual, audio, and text modalities.Our benchmark excels in three key areas: a) Long-Duration, Human-Centric Videos: We carefully selected approximately 1,000 videos from opensource datasets FineVideo based on duration limit and multi-modal information density, focusing on content like lectures, interviews, and vlogs, which contain rich human-centric semantic and contextual attributes.b) Diverse and Challenging Task Scenarios: We have designed six challenging task scenarios, including both Intra-Event and Inter-Event Tasks.c) Rigorous and Comprehensive Quality Assurance Pipelines: We have developed a threestep, semi-automated data quality assurance pipeline to ensure the difficulty and validity of the synthesized questions and answer options.Based on LongInsightBench, we designed a series of experiments.which shows that Omnimodal models(OLMs) still face challenge in tasks requiring precise temporal localization (T-Loc) and long-range causal inference (CE-Caus).Surprisingly, extended experiments reveal the information loss in modal fusion of OLMs, which we called the Fusion Deficit Paradox. 1 * These authors contributed equally.