Route to Reason: Adaptive Routing for LLM and Reasoning Strategy Selection

TL;DR

This paper introduces Route-To-Reason (RTR), a flexible adaptive routing framework that dynamically selects models and reasoning strategies to optimize accuracy and efficiency in reasoning tasks.

Contribution

RTR is a novel unified routing framework that adaptively allocates models and strategies based on task difficulty, reducing costs and improving performance.

Findings

RTR reduces token usage by over 60%.

RTR outperforms single models in accuracy.

RTR demonstrates effectiveness across multiple models and strategies.

Abstract

The inherent capabilities of a language model (LM) and the reasoning strategies it employs jointly determine its performance in reasoning tasks. While test-time scaling is regarded as an effective approach to tackling complex reasoning tasks, it incurs substantial computational costs and often leads to "overthinking", where models become trapped in "thought pitfalls". To address this challenge, we propose Route-To-Reason (RTR), a novel unified routing framework that dynamically allocates both LMs and reasoning strategies according to task difficulty under budget constraints. RTR learns compressed representations of both expert models and reasoning strategies, enabling their joint and adaptive selection at inference time. This method is low-cost, highly flexible, and can be seamlessly extended to arbitrary black-box or white-box models and strategies, achieving true plug-and-play functionality. Extensive experiments across seven open source models and four reasoning strategies demonstrate that RTR achieves an optimal trade-off between accuracy and computational efficiency among all baselines, achieving higher accuracy than the best single model while reducing token usage by over 60%.