[2603.26891] Strategic Candidacy in Generative AI Arenas

Computer Science > Machine Learning arXiv:2603.26891 (cs) [Submitted on 27 Mar 2026] Title:Strategic Candidacy in Generative AI Arenas Authors:Chris Hays, Rachel Li, Bailey Flanigan, Manish Raghavan View a PDF of the paper titled Strategic Candidacy in Generative AI Arenas, by Chris Hays and 3 other authors View PDF HTML (experimental) Abstract:AI arenas, which rank generative models from pairwise preferences of users, are a popular method for measuring the relative performance of models in the course of their organic use. Because rankings are computed from noisy preferences, there is a concern that model producers can exploit this randomness by submitting many models (e.g., multiple variants of essentially the same model) and thereby artificially improve the rank of their top models. This can lead to degradations in the quality, and therefore the usefulness, of the ranking. In this paper, we begin by establishing, both theoretically and in simulations calibrated to data from the platform Arena (formerly LMArena, Chatbot Arena), conditions under which producers can benefit from submitting clones when their goal is to be ranked highly. We then propose a new mechanism for ranking models from pairwise comparisons, called You-Rank-We-Rank (YRWR). It requires that producers submit rankings over their own models and uses these rankings to correct statistical estimates of model quality. We prove that this mechanism is approximately clone-robust, in the sense that a producer canno...