Saved in:
Bibliographic Details
Main Authors: Shirkavand, Reza, Gao, Shangqian, Yu, Peiran, Huang, Heng
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2508.12491
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914169103581184
author Shirkavand, Reza
Gao, Shangqian
Yu, Peiran
Huang, Heng
author_facet Shirkavand, Reza
Gao, Shangqian
Yu, Peiran
Huang, Heng
contents We study cost-aware routing for large language models across diverse and dynamic pools of models. Existing approaches often overlook prompt-specific context, rely on expensive model profiling, assume a fixed set of experts, or use inefficient trial-and-error strategies. We introduce Cost-Spectrum Contrastive Routing (CSCR), a lightweight framework that maps both prompts and models into a shared embedding space to enable fast, cost-sensitive selection. CSCR uses compact, fast-to-compute logit footprints for open-source models and perplexity fingerprints for black-box APIs. A contrastive encoder is trained to favor the cheapest accurate expert within adaptive cost bands. At inference time, routing reduces to a single k-NN lookup via a FAISS index, requiring no retraining when the expert pool changes and enabling microsecond latency. Across multiple benchmarks, CSCR consistently outperforms baselines, improving the accuracy-cost tradeoff by up to 25%, while generalizing robustly to unseen LLMs and out-of-distribution prompts.
format Preprint
id arxiv_https___arxiv_org_abs_2508_12491
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Cost-Aware Contrastive Routing for LLMs
Shirkavand, Reza
Gao, Shangqian
Yu, Peiran
Huang, Heng
Machine Learning
We study cost-aware routing for large language models across diverse and dynamic pools of models. Existing approaches often overlook prompt-specific context, rely on expensive model profiling, assume a fixed set of experts, or use inefficient trial-and-error strategies. We introduce Cost-Spectrum Contrastive Routing (CSCR), a lightweight framework that maps both prompts and models into a shared embedding space to enable fast, cost-sensitive selection. CSCR uses compact, fast-to-compute logit footprints for open-source models and perplexity fingerprints for black-box APIs. A contrastive encoder is trained to favor the cheapest accurate expert within adaptive cost bands. At inference time, routing reduces to a single k-NN lookup via a FAISS index, requiring no retraining when the expert pool changes and enabling microsecond latency. Across multiple benchmarks, CSCR consistently outperforms baselines, improving the accuracy-cost tradeoff by up to 25%, while generalizing robustly to unseen LLMs and out-of-distribution prompts.
title Cost-Aware Contrastive Routing for LLMs
topic Machine Learning
url https://arxiv.org/abs/2508.12491