V-ReasonBench: Toward Unified Reasoning Benchmark Suite for Video Generation Models

arXiv 2025 30.1 method

TLDR

Introduces V-ReasonBench, a benchmark for evaluating video reasoning in generative models across four dimensions using synthetic and real-world sequences.

Reasoning

The paper presents a well-structured benchmark with clear dimensions and evaluation of six models, but its focus is on reasoning evaluation rather than world model development. Strengths include reproducibility and real-world data; weakness is limited novelty in world model concepts.

Read-first score

Read-first score 30.1, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 9.

Recency 8%

86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Methodology quality 25%

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=benchmark,evaluation

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Topical relevance 42%

12.9

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 88.

Keyword Scores

video world model

world model

world dynamics prediction

world simulator

generative world model

interactive world model

model-based reinforcement learning world model

Resources

Paper