Can MiniMax-H3 Reason About the Physical World? An Evaluation of Omni-Modal Generative Model

arXiv 2026 25.4 benchmark

TLDR

Evaluates MiniMax-H3's physical world reasoning via omni-modal tasks combining text, frames, audio, and video; achieves 41.97% success across 517 instances.

Reasoning

The paper introduces a novel multimodal evaluation framework that leverages omni-modal inputs to test physical world reasoning, which is a strength. However, it evaluates only one model and lacks comparisons or baselines in the abstract, limiting evidence of broader significance.

Read-first score

Read-first score 25.4, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 22.

Recency 5%

100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Methodology quality 15%

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

Reproducibility 15%

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

Topical relevance 25%

31.4

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

Graph prestige 15%

5.7

PageRank-style prestige over the local citation graph built from available OpenAlex references. local_pagerank=0.0016

Citation impact 15%

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. cited_by_count=0

Citation velocity 10%

Citation velocity estimates citations per publication-year to reduce old-paper bias. velocity=0.00

Field roles

Frontier

Rank sensitivity

Stability: volatile; rank range: 140.

Keyword Scores

world dynamics prediction

world model

generative world model

video world model

world simulator

interactive world model

model-based reinforcement learning world model

Resources

Paper