Science Utopia? Closed-Loop LLM Simulation of Academic Research Ecosystems

SciUtopia simulates 61 research worlds with 40,000+ LLM-agent researchers, finding rejection-driven resubmission amplifies reviewer burden and inequality emerges endogenously.

SciUtopia is a persistent, closed-loop LLM-agent simulation framework modeling academic ecosystems including research direction choice, collaboration, submission, peer review, resubmission, citation, funding, and attrition with evolving state across simulated years. Across 61 simulation worlds it models over 40,000 researchers from 8,000 institutions, generating roughly 400,000 publication decisions and 1.2 million LLM-generated peer reviews. Findings include that rejection-driven resubmission substantially amplifies reviewer burden beyond population growth, cautious exploration balances citation impact with career success and topic diversity, and resource inequality emerges even without detectable cumulative advantage from narrowly winning early funding. The framework supports matched counterfactual experiments and targeted interventions, with code available on GitHub.