Tandem: Riding Together with Large and Small Language Models for Efficient Reasoning

TL;DR

Tandem is a collaborative framework that combines large and small language models to perform reasoning tasks more efficiently, reducing computational costs by about 40% while maintaining high performance.

Contribution

The paper introduces a novel LLM-SLM collaboration approach with a cost-aware termination mechanism for efficient reasoning, with code available online.

Findings

Reduces computational costs by approximately 40% compared to standalone LLM reasoning.

Achieves superior or competitive performance on mathematical reasoning and code generation benchmarks.

Sufficiency classifier transfers effectively across different domains without retraining.

Abstract

Recent advancements in large language models (LLMs) have catalyzed the rise of reasoning-intensive inference paradigms, where models perform explicit step-by-step reasoning before generating final answers. While such approaches improve answer quality and interpretability, they incur substantial computational overhead due to the prolonged generation sequences. In this paper, we propose Tandem, a novel collaborative framework that synergizes large and small language models (LLMs and SLMs) to achieve high-quality reasoning with significantly reduced computational cost. Specifically, the LLM serves as a strategic coordinator, efficiently generating a compact set of critical reasoning insights. These insights are then used to guide a smaller, more efficient SLM in executing the full reasoning process and delivering the final response. To balance efficiency and reliability, Tandem introduces a cost-aware termination mechanism that adaptively determines when sufficient reasoning guidance has been accumulated, enabling early stopping of the LLM's generation. Experiments on mathematical reasoning and code generation benchmarks demonstrate that Tandem reduces computational costs by approximately 40% compared to standalone LLM reasoning, while achieving superior or competitive performance. Furthermore, the sufficiency classifier trained on one domain transfers effectively to others without retraining. The code is available at: https://github.com/Applied-Machine-Learning-Lab/ACL2026_Tandem.