ORI: O Routing Intelligence

TL;DR

ORI is a dynamic routing framework that intelligently directs queries to the most suitable large language model, enhancing accuracy and efficiency across diverse tasks.

Contribution

This paper introduces ORI, a novel adaptive routing system that leverages multiple LLMs to improve task-specific performance and scalability.

Findings

Outperforms individual models on MMLU and MuSR benchmarks

Achieves up to 2.7 points improvement on MMLU

Maintains efficiency while enhancing accuracy

Abstract

Single large language models (LLMs) often fall short when faced with the ever-growing range of tasks, making a single-model approach insufficient. We address this challenge by proposing ORI (O Routing Intelligence), a dynamic framework that leverages a set of LLMs. By intelligently routing incoming queries to the most suitable model, ORI not only improves task-specific accuracy, but also maintains efficiency. Comprehensive evaluations across diverse benchmarks demonstrate consistent accuracy gains while controlling computational overhead. By intelligently routing queries, ORI outperforms the strongest individual models by up to 2.7 points on MMLU and 1.8 points on MuSR, ties the top performance on ARC, and on BBH. These results underscore the benefits of a multi-model strategy and demonstrate how ORI's adaptive architecture can more effectively handle diverse tasks, offering a scalable, high-performance solution for a system of multiple large language models.