A Very Big Video Reasoning Suite

TL;DR

This paper introduces the Very Big Video Reasoning (VBVR) Dataset and VBVR-Bench, large-scale resources for evaluating and understanding video reasoning capabilities, enabling systematic scaling studies and fostering progress in generalizable video reasoning.

Contribution

The paper presents the first large-scale video reasoning dataset and benchmark, facilitating systematic evaluation and scaling studies of video reasoning models.

Findings

Early signs of emergent generalization to unseen reasoning tasks.

The VBVR suite enables reproducible and interpretable diagnosis of video reasoning.

The dataset spans over one million video clips across 200 reasoning tasks.

Abstract

Rapid progress in video models has largely focused on visual quality, leaving their reasoning capabilities underexplored. Video reasoning grounds intelligence in spatiotemporally consistent visual environments that go beyond what text can naturally capture, enabling intuitive reasoning over spatiotemporal structure such as continuity, interaction, and causality. However, systematically studying video reasoning and its scaling behavior is hindered by the lack of large-scale training data. To address this gap, we introduce the Very Big Video Reasoning (VBVR) Dataset, an unprecedentedly large-scale resource spanning 200 curated reasoning tasks following a principled taxonomy and over one million video clips, approximately three orders of magnitude larger than existing datasets. We further present VBVR-Bench, a verifiable evaluation framework that moves beyond model-based judging by incorporating rule-based, human-aligned scorers, enabling reproducible and interpretable diagnosis of video reasoning capabilities. Leveraging the VBVR suite, we conduct one of the first large-scale scaling studies of video reasoning and observe early signs of emergent generalization to unseen reasoning tasks. Together, VBVR lays a foundation for the next stage of research in generalizable video reasoning. The data, benchmark toolkit, and models are publicly available at https://video-reason.com/ .