Impacts of Aggregation on Model Diversity and Consumer Utility
2026-02-26 • Computer Science and Game Theory
Computer Science and Game TheoryComputers and Society
AI summaryⓘ
The authors look at how users can choose from a variety of AI models with different strengths to get better results than using just one model all the time. They study how current evaluation methods push creators to make models that are too similar, which isn't the best for users. To fix this, they suggest a new scoring method called weighted winrate that encourages making diverse models specialized in different tasks. They also show that this method helps both in theory and in real benchmark tests.
AI marketplacemodel selectionwinrateweighted winratemodel specializationconsumer welfaremodel evaluationbenchmark datasetsincentive mechanismslarge language models
Authors
Kate Donahue, Manish Raghavan
Abstract
Consider a marketplace of AI tools, each with slightly different strengths and weaknesses. By selecting the right model for the task at hand, a user can do better than simply committing to a single model for everything. Routers operate under a similar principle, where sophisticated model selection can increase overall performance. However, aggregation is often noisy, reflecting in imperfect user choices or routing decisions. This leads to two main questions: first, what does a "healthy marketplace" of models look like for maximizing consumer utility? Secondly, how can we incentivize producers to create such models? Here, we study two types of model changes: market entry (where an entirely new model is created and added to the set of available models), and model replacement (where an existing model has its strengths and weaknesses changed). We show that winrate, a standard benchmark in LLM evaluation, can incentivize model creators to homogenize for both types of model changes, reducing consumer welfare. We propose a new mechanism, weighted winrate, which rewards models for answers that are higher quality, and show that it provably improves incentives for producers to specialize and increases consumer welfare. We conclude by demonstrating that our theoretical results generalize to empirical benchmark datasets and discussing implications for evaluation design.