English

ZeroSumEval: An Extensible Framework For Scaling LLM Evaluation with Inter-Model Competition

Computation and Language 2025-04-18 v2 Artificial Intelligence

Abstract

We introduce ZeroSumEval, a dynamic, competition-based, and evolving evaluation framework for Large Language Models (LLMs) that leverages competitive games. ZeroSumEval encompasses a diverse suite of games, including security challenges (Capture the Flag), classic board games (chess), and knowledge tests (MathQuiz). These games are designed to evaluate a range of capabilities such as strategic reasoning, planning, knowledge application, safety, and adaptability. Building upon recent studies that highlight the effectiveness of game-based evaluations for LLMs, ZeroSumEval enhances these approaches by providing a standardized and extensible framework for easily implementing games and leverages DSPy to provide a better abstraction for LLM player strategies.

Keywords

Cite

@article{arxiv.2503.10673,
  title  = {ZeroSumEval: An Extensible Framework For Scaling LLM Evaluation with Inter-Model Competition},
  author = {Hisham A. Alyahya and Haidar Khan and Yazeed Alnumay and M Saiful Bari and Bülent Yener},
  journal= {arXiv preprint arXiv:2503.10673},
  year   = {2025}
}
R2 v1 2026-06-28T22:19:31.585Z