mlsanityData to 14 Oct 2026

Energy-Aware Routing to Large Reasoning Models

arXiv 2601.00823 ↗First posted 23 Dec 2025cs.AI cs.IT cs.SYcited by 0

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Large reasoning models (LRMs) have heterogeneous inference energy costs based on which model is used and how much it reasons. To reduce energy, it is important to choose the right LRM and operate it in the right way. As a result, the performance of systems that dispatch tasks to different individual LRMs depend on the balance between mean energy provisioning and stochastic fluctuations. The critical regime is the unique operating point at which neither auxiliary energy nor baseline energy is systematically wasted. Increasing baseline supply shifts the system toward persistent over-supply and baseline-energy waste, while reducing supply induces persistent reliance on auxiliary energy. Yet in this regime, performance remains volatility-limited and so a second-order characterization provides further insights that we develop. Here, performance is governed by how variability is absorbed across time, models, and execution choices. This perspective highlights variance-aware routing and dispatch as a principled design axis, and provides a theoretical basis for developing energy-aware model routing policies. Routing behavior is characterized when dispatch policies are based on training-compute and inference-compute scaling laws for LRMs.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 20 of its 41 references that are in the map. First posted on arXiv 23 Dec 2025.

Builds on

20 of 41 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025A Theory of Inference Compute Scaling: Reasoning through Directed Stochastic Skill Search10 Jun

Other references · 19

  1. 2020Scaling Laws for Neural Language Modelscited by 6,586
  2. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,944
  3. 2022Training Compute-Optimal Large Language Modelscited by 3,019
  4. 2023FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performancecited by 853
  5. 2023A Simple and Effective Pruning Approach for Large Language Modelscited by 800
  6. 2023Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizescited by 737
  7. 2021Explaining Neural Scaling Lawscited by 363
  8. 2022Estimating the Carbon Footprint of BLOOM, a 176B Parameter Language Modelcited by 265
  9. 2023Large Language Model Routing with Benchmark Datasetscited by 172
  10. 2024Compact Language Models via Pruning and Knowledge Distillationcited by 148
  11. 2024Multi-Step Reasoning with Large Language Models, a Surveycited by 143
  12. 2025How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM Inferencecited by 76
7 more
  1. 2024The Price of Prompting: Profiling Energy Use in Large Language Models Inferencecited by 23
  2. 2025EcoServe: Designing Carbon-Aware AI Inference Systemscited by 19
  3. 2023Clover: Toward Sustainable AI with Carbon-Aware Machine Learning Inference Servicecited by 17
  4. 2024Measuring the Energy Consumption and Efficiency of Deep Neural Networks: An Empirical Analysis and Design Recommendationscited by 12
  5. 2025Foundation Models for Discovery and Exploration in Chemical Spacecited by 6
  6. 2025Carbon- and Precedence-Aware Scheduling for Data Processing Clusterscited by 2
  7. 2017Stochastic Routing and Scheduling Policies for Energy Harvesting Communication Networkscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Measured Joules, Learned Routes: Learning to Route for Energy-Efficient LLM Serving
  2. 2026Reasoning Is Not Free: Robust Adaptive Cost-Efficient Routing for LLM-as-a-Judge
  3. 2026Measuring the Inference Energy Efficiency of Open-Weight LLMs across Benchmarks, Reasoning Modes, and Hardware
  4. 2026Policy-Guided Stepwise Model Routing for Cost-Effective Reasoning
  5. 2026Rethinking Stepwise Model Routing: A Cost-Efficient Table Reasoning Perspective
  6. 2026From Sampled Outcomes to Capability Distributions: Rethinking Supervision for LLM Routing
  7. 2026Models Under SCOPE: Scalable and Controllable Routing via Pre-hoc Reasoning
  8. 2025Confidence-Guided Stepwise Model Routing for Cost-Efficient Reasoning