Video-DeepResearch：迈向下一代多模态深度研究智能体

摘要

我们提出Video-DeepResearch（Video-DR），将多模态智能体从静态图像扩展到连续视频流——这一设定要求密集的时空定位能力与开放网络探索能力的结合。初步评估揭示了当前模型存在的两个关键瓶颈：（1）模态偏差，即智能体绕过视觉工具而倾向于文本搜索；（2）参数化知识泄漏，即模型依赖内部记忆而非真正基于工具的增强执行。为应对这些挑战，我们提出Video-DR，其核心是一个解耦的感知-探索流水线，并采用分阶段工具解锁策略，强制模型在网络检索之前进行全面的跨帧视觉定位。我们的框架采用两阶段训练方案：监督微调后接组相对策略优化（GRPO），从而实现自主探索，突破模仿学习的性能天花板。此外，我们构建了Video-DR-Bench，一个包含200个复杂多跳视觉问答（VQA）实例的人机协同基准。实验结果表明，我们的Video-DeepResearch-35B-A3B以64.0%的平均准确率取得了新的最先进水平，超越专有模型Claude-4.5-Sonnet（59.0%）5.0个百分点，并显著优于GPT-5（52.5%）和Gemini 2.5 Pro（57.5%）。30B-A3B变体达到59.3%的准确率，与Claude-4.5-Sonnet相当，证明了我们的训练范式即使在紧凑规模下同样有效。代码：https://github.com/Osilly/Vision-DeepResearch。

English

We introduce Video-DeepResearch (Video-DR), extending multimodal agents from static images to continuous video streams, a setting that demands dense spatiotemporal grounding coupled with open-web exploration. Preliminary evaluations reveal two critical bottlenecks in current models: (1) modality bias, where agents bypass visual tools in favor of textual search, and (2) parametric knowledge leakage, where models rely on internal memory rather than genuine tool-augmented execution. To address these challenges, we propose Video-DR, featuring a decoupled perception-exploration pipeline with stage-wise tool unlocking that compels exhaustive cross-frame visual grounding prior to web retrieval. Our framework adopts a two-stage training recipe: supervised fine-tuning followed by Group Relative Policy Optimization (GRPO), enabling autonomous exploration that breaks the imitation-learning ceiling. Furthermore, we curate Video-DR-Bench, a human-AI collaborative benchmark comprising 200 complex, multi-hop VQA instances. Empirical results demonstrate that our Video-DeepResearch-35B-A3B establishes a new state-of-the-art of 64.0% average accuracy, surpassing proprietary Claude-4.5-Sonnet (59.0%) by 5.0 points and significantly outperforming GPT-5 (52.5%) and Gemini 2.5 Pro (57.5%). The 30B-A3B variant achieves 59.3%, competitive with Claude-4.5-Sonnet and demonstrating the effectiveness of our training paradigm even at compact scale. Code: https://github.com/Osilly/Vision-DeepResearch.