JANUS: Enhancing Long-Horizon Autonomous Research with Co-Evolving World Models

arXiv 2026 28.2 method, application

TLDR

JANUS uses a co-evolving world model to forecast and reflect in a Predict-Verify-Evolve loop, improving long-horizon autonomous research on benchmarks and real tasks.

Reasoning

The paper proposes a novel framework that co-evolves a research agent and world model to reduce feedback delays, with strong empirical results on MLS-Bench-Lite and real research tasks. However, the abstract lacks details on limitations, scalability, and the exact nature of real-world tasks, and several world-model-related keywords are only weakly supported by the visible text.

Read-first score

Read-first score 28.2, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 29.

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=baseline,evaluation,validation

Topical relevance 25%

41.4

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

Reproducibility 15%

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

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: 188.

Keyword Scores

world model

world simulator

world dynamics prediction

model-based reinforcement learning world model

generative world model

interactive world model

video world model

Resources

Paper