Search papers, labs, and topics across Lattice.
The paper introduces LongVideo-R1, a multimodal large language model (MLLM) agent designed for efficient long video understanding under low computational budgets by actively navigating video context. LongVideo-R1 employs a reasoning module that uses high-level visual cues to infer the most informative video clips, enabling iterative refinement and early halting of exploration. Fine-tuning the Qwen-3-8B model using supervised fine-tuning (SFT) followed by reinforcement learning (RL) with a custom reward function, LongVideo-R1 achieves a superior trade-off between QA accuracy and efficiency on long video benchmarks.
Achieve state-of-the-art long video QA accuracy with significantly reduced computational cost by actively reasoning about which video segments are most relevant.
This paper addresses the critical and underexplored challenge of long video understanding with low computational budgets. We propose LongVideo-R1, an active, reasoning-equipped multimodal large language model (MLLM) agent designed for efficient video context navigation, avoiding the redundancy of exhaustive search. At the core of LongVideo-R1 lies a reasoning module that leverages high-level visual cues to infer the most informative video clip for subsequent processing. During inference, the agent initiates traversal from top-level visual summaries and iteratively refines its focus, immediately halting the exploration process upon acquiring sufficient knowledge to answer the query. To facilitate training, we first extract hierarchical video captions from CGBench, a video corpus with grounding annotations, and guide GPT-5 to generate 33K high-quality chain-of-thought-with-tool trajectories. The LongVideo-R1 agent is fine-tuned upon the Qwen-3-8B model through a two-stage paradigm: supervised fine-tuning (SFT) followed by reinforcement learning (RL), where RL employs a specifically designed reward function to maximize selective and efficient clip navigation. Experiments on multiple long video benchmarks validate the effectiveness of name, which enjoys superior tradeoff between QA accuracy and efficiency. All curated data and source code are provided in the supplementary material and will be made publicly available. Code and data are available at: https://github.com/qiujihao19/LongVideo-R1
