Video Models Start to Solve Chess, Maze, Sudoku, Mental Rotation, and Raven' Matrices

arXiv 2025 37.2 method

TLDR

Video generation models show reasoning abilities on chess, mazes, Sudoku, mental rotation, and Raven's Matrices, with a scalable Task Pair evaluation paradigm and VMEvalKit framework.

Reasoning

The paper provides a strong empirical paradigm with many models and tasks, and validates automated evaluation against human judgment. However, the abstract focuses on video models as reasoners, not world models, so world-model-related keywords have low direct relevance.

Read-first score

Read-first score 37.2, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 10.

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=evaluation,experiment,result

Reproducibility 25%

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=code,github

Topical relevance 42%

14.3

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: 453.

Keyword Scores

video world model

world model

world simulator

generative world model

interactive world model

world dynamics prediction

model-based reinforcement learning world model

Resources

Paper