LVOmniBench: Pioneering Long Audio-Video Understanding Evaluation for Omnimodal LLMs
This addresses a critical gap for real-world applications where videos are typically long, though it is incremental as it focuses on evaluation rather than model development.
The authors tackled the lack of evaluation for long audio-video understanding in omnimodal LLMs by introducing LVOmniBench, a benchmark with 275 videos (10-90 minutes) and 1,014 QA pairs, finding that current models struggle with accuracies below 35% for open-source models and up to 65% for Gemini 3 Pro.
Recent advancements in omnimodal large language models (OmniLLMs) have significantly improved the comprehension of audio and video inputs. However, current evaluations primarily focus on short audio and video clips ranging from 10 seconds to 5 minutes, failing to reflect the demands of real-world applications, where videos typically run for tens of minutes. To address this critical gap, we introduce LVOmniBench, a new benchmark designed specifically for the cross-modal comprehension of long-form audio and video. This dataset comprises high-quality videos sourced from open platforms that feature rich audio-visual dynamics. Through rigorous manual selection and annotation, LVOmniBench comprises 275 videos, ranging in duration from 10 to 90 minutes, and 1,014 question-answer (QA) pairs. LVOmniBench aims to rigorously evaluate the capabilities of OmniLLMs across domains, including long-term memory, temporal localization, fine-grained understanding, and multimodal perception. Our extensive evaluation reveals that current OmniLLMs encounter significant challenges when processing extended audio-visual inputs. Open-source models generally achieve accuracies below 35%, whereas the Gemini 3 Pro reaches a peak accuracy of approximately 65%. We anticipate that this dataset, along with our empirical findings, will stimulate further research and the development of advanced models capable of resolving complex cross-modal understanding problems within long-form audio-visual contexts.