VideoSearcher: Empowering Video Deep Research with Multi-Tool Agentic Reasoning via Reinforcement Learning
This work addresses the lack of video-centric search agents for open-world evidence exploration, enabling Vision-Language Models to perform complex video reasoning tasks.
VideoSearcher introduces a multi-tool agentic framework for Video Deep Research (VDR), unifying temporal localization, spatial focusing, and multimodal search. It achieves significant improvements over prior open-source baselines on search-oriented and multimodal understanding benchmarks.
Video understanding is moving beyond closed-context perception toward open-world evidence exploration, a paradigm formalized as Video Deep Research (VDR). However, existing multimodal search agents primarily target static images, and the current VDR benchmark relies on text-centric retrieval that discards crucial visual information. To address these limitations, we propose VideoSearcher, a closed-loop agentic framework that empowers Vision-Language Models with multi-tool reasoning for VDR. VideoSearcher unifies temporal localization, spatial focusing, and multimodal search within a single reasoning trajectory, enabling agents to progressively ground visual clues, retrieve relevant evidence, and synthesize answers. To optimize knowledge-intensive reasoning trajectories, we propose Bi-branch Sequence Policy Optimization (BiSPO), a reinforcement learning algorithm that decouples tool-invocation optimization from answer-accuracy optimization. This design provides stable learning signals for both evidence-grounded reasoning and purposeful tool use. Furthermore, we construct VideoSearch-QA, the first benchmark designed to evaluate open-world video information grounding and multimodal search-based reasoning. Extensive experiments demonstrate that VideoSearcher significantly outperforms prior open-source agentic baselines across various search-oriented and multimodal understanding benchmarks.