# Beyond Test-Time Compute Strategies: Advocating Energy-per-Token in LLM Inference

arXiv: https://arxiv.org/abs/2603.20224 · HTML: https://mlsanity.com/paper/2603.20224/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 23 of its 38 references that are in the map.
- First posted on arXiv 4 Mar 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 23 of 23)

- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [Let's Verify Step by Step](https://mlsanity.com/paper/2305.20050/) (2023)
- [Training Compute-Optimal Large Language Models](https://mlsanity.com/paper/2203.15556/) (2022)
- [Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding](https://mlsanity.com/paper/1510.00149/) (2015)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotations](https://mlsanity.com/paper/2312.08935/) (2023)
- [Solving math word problems with process- and outcome-based feedback](https://mlsanity.com/paper/2211.14275/) (2022)
- [A Simple and Effective Pruning Approach for Large Language Models](https://mlsanity.com/paper/2306.11695/) (2023)
- [Alphazero-like Tree-Search can Guide Large Language Model Decoding and Training](https://mlsanity.com/paper/2309.17179/) (2023)
- [Chain-of-Thought Reasoning Without Prompting](https://mlsanity.com/paper/2402.10200/) (2024)
- [DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search](https://mlsanity.com/paper/2408.08152/) (2024)
- [Beyond Chinchilla-Optimal: Accounting for Inference in Language Model Scaling Laws](https://mlsanity.com/paper/2401.00448/) (2023)
- [DynamoLLM: Designing LLM Inference Clusters for Performance and Energy Efficiency](https://mlsanity.com/paper/2408.00741/) (2024)
- [Compressing LLMs: The Truth is Rarely Pure and Never Simple](https://mlsanity.com/paper/2310.01382/) (2023)
- [Counting Carbon: A Survey of Factors Influencing the Emissions of Machine Learning](https://mlsanity.com/paper/2302.08476/) (2023)
- [Offline Energy-Optimal LLM Serving: Workload-Based Energy Models for LLM Inference on Heterogeneous Systems](https://mlsanity.com/paper/2407.04014/) (2024)
- [From Efficiency Gains to Rebound Effects: The Problem of Jevons' Paradox in AI's Polarized Environmental Debate](https://mlsanity.com/paper/2501.16548/) (2025)
- [Toward Sustainable GenAI using Generation Directives for Carbon-Friendly Large Language Model Inference](https://mlsanity.com/paper/2403.12900/) (2024)

## Built on by (most cited first, 5 of 5)

- [Rethinking Model Efficiency: Multi-Agent Inference with Large Models](https://mlsanity.com/paper/2604.04929/) (2026)
- [The Shadow Price of Reasoning: Economic Perspective on Optimal Budget Allocation for LLMs](https://mlsanity.com/paper/2606.03092/) (2026)
- [Characterization of Request and Token Energy Costs for LLM Inference Workloads on GPU Platforms](https://mlsanity.com/paper/2608.28044/) (2026)
- [KAIROS: Stateful, Context-Aware Power-Efficient Agentic Inference Serving](https://mlsanity.com/paper/2604.16682/) (2026)
- [GreenBench: Benchmarking Energy Efficiency and Carbon Footprint of Open-Source LLM Inference on Apple Silicon](https://mlsanity.com/paper/2608.28667/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) demonstrate exceptional performance across diverse tasks but come with substantial energy and computational costs, particularly in request-heavy scenarios. In many real-world applications, the full scale and capabilities of LLMs are often unnecessary, as Small Language Models (SLMs) can provide accurate responses for simpler text generation tasks. When enhanced with advanced reasoning strategies, such as Chain-of-Thought (CoT) prompting or Majority Voting, SLMs can approach the performance of larger models while reducing overall computational requirements. However, these strategies can also introduce additional energy costs, creating an energy-accuracy trade-off. Our analysis examines these trade-offs in test-time compute strategies for smaller models compared to larger ones, using the MMLU benchmark. Additionally, we explore the input-output token dynamics of transformer architectures, which result in nonlinear hardware energy operation curves for LLMs. To bridge AI research with its physical impact, we propose textitenergy efficiency metrics, including Energy-per-Token, as complements to traditional accuracy benchmarks. Beyond model selection, we propose controlled reasoning in CoT token generation, using operating curves to regulate reasoning depth dynamically. This vision integrates a energy-aware routing mechanism, ensuring that model selection and inference strategies balance accuracy for sustainable AI deployment.