Video-o3: Native Interleaved Clue Seeking for Long Video Multi-Hop Reasoning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zeng, Xiangyu, Zhang, Zhiqiu, Zhu, Yuhan, Li, Xinhao, Wang, Zikang, Ma, Changlian, Zhang, Qingyu, Huang, Zizheng, Ouyang, Kun, Jiang, Tianxiang, Yan, Ziang, Wang, Yi, Zhang, Hongjie, Wang, Yali, Wang, Limin
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916034366144512
author Zeng, Xiangyu
Zhang, Zhiqiu
Zhu, Yuhan
Li, Xinhao
Wang, Zikang
Ma, Changlian
Zhang, Qingyu
Huang, Zizheng
Ouyang, Kun
Jiang, Tianxiang
Yan, Ziang
Wang, Yi
Zhang, Hongjie
Wang, Yali
Wang, Limin
author_facet Zeng, Xiangyu
Zhang, Zhiqiu
Zhu, Yuhan
Li, Xinhao
Wang, Zikang
Ma, Changlian
Zhang, Qingyu
Huang, Zizheng
Ouyang, Kun
Jiang, Tianxiang
Yan, Ziang
Wang, Yi
Zhang, Hongjie
Wang, Yali
Wang, Limin
contents Existing multimodal large language models for long-video understanding predominantly rely on uniform sampling and single-turn inference, limiting their ability to identify sparse yet critical evidence amid extensive redundancy. We introduce Video-o3, a novel framework that supports iterative discovery of salient visual clues, fine-grained inspection of key segments, and adaptive termination once sufficient evidence is acquired. Technically, we address two core challenges in interleaved tool invocation. First, to mitigate attention dispersion induced by the heterogeneity of reasoning and tool-calling, we propose Task-Decoupled Attention Masking, which isolates per-step concentration while preserving shared global context. Second, to control context length growth in multi-turn interactions, we introduce a Verifiable Trajectory-Guided Reward that balances exploration coverage with reasoning efficiency. To support training at scale, we further develop a data synthesis pipeline and construct Seeker-173K, comprising 173K high-quality tool-interaction trajectories for effective supervised and reinforcement learning. Extensive experiments show that Video-o3 substantially outperforms state-of-the-art methods, achieving 72.1% accuracy on MLVU and 46.5% on Video-Holmes. These results demonstrate Video-o3's strong multi-hop evidence-seeking and reasoning capabilities, and validate the effectiveness of native tool invocation in long-video scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2601_23224
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Video-o3: Native Interleaved Clue Seeking for Long Video Multi-Hop Reasoning
Zeng, Xiangyu
Zhang, Zhiqiu
Zhu, Yuhan
Li, Xinhao
Wang, Zikang
Ma, Changlian
Zhang, Qingyu
Huang, Zizheng
Ouyang, Kun
Jiang, Tianxiang
Yan, Ziang
Wang, Yi
Zhang, Hongjie
Wang, Yali
Wang, Limin
Computer Vision and Pattern Recognition
Existing multimodal large language models for long-video understanding predominantly rely on uniform sampling and single-turn inference, limiting their ability to identify sparse yet critical evidence amid extensive redundancy. We introduce Video-o3, a novel framework that supports iterative discovery of salient visual clues, fine-grained inspection of key segments, and adaptive termination once sufficient evidence is acquired. Technically, we address two core challenges in interleaved tool invocation. First, to mitigate attention dispersion induced by the heterogeneity of reasoning and tool-calling, we propose Task-Decoupled Attention Masking, which isolates per-step concentration while preserving shared global context. Second, to control context length growth in multi-turn interactions, we introduce a Verifiable Trajectory-Guided Reward that balances exploration coverage with reasoning efficiency. To support training at scale, we further develop a data synthesis pipeline and construct Seeker-173K, comprising 173K high-quality tool-interaction trajectories for effective supervised and reinforcement learning. Extensive experiments show that Video-o3 substantially outperforms state-of-the-art methods, achieving 72.1% accuracy on MLVU and 46.5% on Video-Holmes. These results demonstrate Video-o3's strong multi-hop evidence-seeking and reasoning capabilities, and validate the effectiveness of native tool invocation in long-video scenarios.
title Video-o3: Native Interleaved Clue Seeking for Long Video Multi-Hop Reasoning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.23224