minWM: A Full-Stack Open-Source Framework for Real-Time Interactive Video World Models

arXiv 2026 71.9 method, system

TLDR

minWM is an open-source framework for building real-time interactive video world models from video diffusion models via controllable fine-tuning and distillation.

Reasoning

The paper presents a comprehensive pipeline for converting video diffusion models into interactive world models, with modular design and support for multiple backbones. However, the abstract lacks explicit experimental results or real-world benchmarks, making it unclear how well the framework performs in practice.

Read-first score

Read-first score 71.9, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 55.

Recency 6%

100

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

Citation impact 18%

88.2

Uses OpenAlex-shaped citation metadata as a bibliometric attention signal, separate from paper quality. citation_normalized_percentile=0.88161174

Reproducibility 18%

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

Topical relevance 29%

78.6

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

Methodology quality 18%

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

Citation velocity 12%

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

Field roles

FoundationFrontierBridgeMethodology anchorReproducibility anchor

Rank sensitivity

Stability: volatile; rank range: 391.

Keyword Scores

world model

interactive world model

video world model

generative world model

world simulator

world dynamics prediction

model-based reinforcement learning world model

Deep Analysis

Innovations

Full-stack open-source framework for building real-time interactive video world models
End-to-end pipeline converting bidirectional T2V/TI2V video foundation models into camera-controllable few-step autoregressive world models
Causal Forcing / Causal Forcing++ pipeline including AR diffusion training, causal ODE or causal consistency distillation, and asymmetric DMD
Modular and architecture-extensible instantiation on representative open backbones (Wan2.1-T2V-1.3B and HY1.5-TI2V-8B)
Support for adapting existing video world models (e.g., HY-WorldPlay) to new data distributions, training recipes, and latency targets

Methodology

minWM provides an end-to-end pipeline that first fine-tunes a bidirectional video diffusion model with camera control, then applies the Causal Forcing/Causal Forcing++ pipeline (including AR diffusion training, causal ODE or causal consistency distillation, and asymmetric DMD) to distill it into a few-step autoregressive generator for low-latency rollout. The framework is modular and architecture-extensible, instantiated on Wan2.1-T2V-1.3B and HY1.5-TI2V-8B.

Key Results

The paper provides practical ablations on camera trajectory quality, controllability training steps, and minimal batch-size requirements, but no specific quantitative results are mentioned in the abstract.

Resources

Paper Code