An Initial Exploration of Contrastive Prompt Tuning to Generate Energy-Efficient Code

Sophie Weidmann; Fernando Castor

cs.LGMar 3, 2026

An Initial Exploration of Contrastive Prompt Tuning to Generate Energy-Efficient Code

Authors: Sophie Weidmann, Fernando Castor

Organizations: University of Twente, Enschede, The Netherlands

Abstract

Although LLMs are capable of generating functionally correct code, they also tend to produce less energy-efficient code in comparison to human-written solutions. As these inefficiencies lead to higher computational overhead, they are in direct conflict with Green Software Development (GSD) efforts, which aim to reduce the energy consumption of code. To support these efforts, this study aims to investigate whether and how LLMs can be optimized to promote the generation of energy-efficient code. To this end, we employ Contrastive Prompt Tuning (CPT). CPT combines Contrastive Learning techniques, which help the model to distinguish between efficient and inefficient code, and Prompt Tuning, a Parameter-Efficient Fine Tuning (PEFT) approach that requires only a fraction of the cost of traditional fine tuning. This study evaluates CPT on Python, Java and C++ coding problems across three different models to provide a comprehensive evaluation. The method achieves consistent improvements in code accuracy for two models but efficiency gains vary by model, language and task complexity, indicating that improvements are not uniformly reliable.

Figures & tables

Figure 1. In prompt tuning, the soft prompt is optimized through backpropagation and the model remains frozen. Input consists of learnable soft prompt and standard prompt.

Figure 2. Overview of the experimental design and configuration space for the training and generation phases.

Abbrevation	Model Type	Model Name	Parameters
general	General-Purpose	Phi-3 Mini 3 3 3 https://huggingface.co/microsoft/Phi-3-mini-4k-instruct	$\sim$ 3.8B
tuned	Code-Tuned	CodeLlama-7B 4 4 4 https://huggingface.co/codellama/CodeLlama-7b-hf	$\sim$ 7B
first	Code-First	DeepSeek Coder 5 5 5 https://huggingface.co/deepseek-ai/deepseek-coder-7b-instruct	$\sim$ 7B

Table 1. Selected models for the experiment and evaluation.

Training Sets
Dataset	Language(s)	Size
Python	Python	3,548 entries
C ++	C ++	3,921 entries
Both	Python, C ++	7,469 entries
Benchmark Sets
Dataset	Language(s)	Size

Table 2. Overview of training and benchmark datasets used in this work.

Component	Specification
Device	MacBook Pro (15-inch, 2016)
Processor	2.7 GHz Quad-Core Intel Core i7
Memory	16 GB 2133 MHz LPDDR3
Graphics	Intel HD Graphics 530 (1536 MB)
Storage	512 GB SSD
OS	macOS Montery v12.7.6

Table 3. Hardware setup used for evaluation.

Table 4. Pass@1 results (in %) for all variants of the three models. Percentages are reported for the 648 tasks, followed by results on HumanEval-X and LeetCode, respectively 164 and 52 tasks per language: C ++ , Java, Python. The best-performing non-baseline, non-instruction variant for each model is indicated in green and the results pertain to the epoch with the best training results (“ ep ”) We report the baseline model accuracy for the original version, with no CPT.

Code Generation

Code generation studies methods for turning natural-language, visual, or formal specifications into executable programs and code edits.

See all 313 papers

Improving the Energy-Efficiency of the Code Generated by LLMs through Effective Prompting
Oct 1, 2026Ritika Rekhi, Bing Zhang, Md Arman Islam +7LLM Code Generation Code Generation
University at Buffalo, Buffalo, NY, USA · National Center for Supercomputing Applications, Urbana, IL, USA · IBM Research, Yorktown Heights, NY, USA
Beyond the Need for Speed: Energy-Aware Code Generation via Simulation-Guided Reinforcement Learning
Jul 6, 2026Saurabhsingh Rajput, Tushar SharmaRL for Code Generation Reinforcement Learning
Dalhousie University, Halifax, Canada
Green Code or Costly Code? Evaluating the Energy Costs of Iterative LLM-Based Code Optimizations
Oct 23, 2026Sadra Emami, Benjamin Weigell, Mohammad Reza Mousavi +3Large Language Model-Guided Optimization Code Generation
University of Augsburg, Augsburg, Germany · King’s College London, London, United Kingdom

Oct 1, 2026J/K move · Enter open · S save

Improving the Energy-Efficiency of the Code Generated by LLMs through Effective Prompting

Ritika Rekhi, Bing Zhang, Md Arman Islam, Jaya Krishna Pasham+6

University at Buffalo, Buffalo, NY, USA · National Center for Supercomputing Applications, Urbana, IL, USA · IBM Research, Yorktown Heights, NY, USA

LLM Code Generation Code Generation LLM Prompting Code Generation Evaluation

Open paper AskPDF arXiv

As AI-assisted programming becomes increasingly mainstream, the environmental impact of AI-generated software has emerged as an important consideration. This motivates evaluating LLM-generated code beyond functional correctness by considering execution efficiency and energy consumption. However, despite substantial advances in code generation, frontier LLMs are rarely evaluated based on the energy efficiency of the code they produce. In this work, we conduct a comprehensive evaluation of 21 prompting strategies for energy-efficient code generation and identify 8 strategies for evaluation across 10 widely used open-weight and proprietary LLMs. We evaluate their effectiveness for both Python and C++ code generation relative to a baseline prompt. Across the evaluated models, the selected prompting strategies achieved energy reductions of up to 25% for Python and 17% for C++ code generation. At the model level, Python energy reductions reached up to 50% for Granite-4.0-H-Small, 39% for Claude 4.5 Haiku, and 28% for MiniMax M3, while C++ reductions reached up to 57% for Granite-4.0-H-Small and 6.4% for Qwen3-Coder-480B-A35B-Instruct. These results demonstrate that prompting strategies can substantially influence the energy consumption of LLM-generated code, although their effectiveness varies across models and programming languages. Our findings highlight the importance of incorporating energy efficiency into the evaluation and optimization of LLM-based code generation and provide practical insights into designing prompts for more sustainable AI-assisted programming.

Abstract

Code Generation

Explore similar work

Improving the Energy-Efficiency of the Code Generated by LLMs through Effective Prompting

Beyond the Need for Speed: Energy-Aware Code Generation via Simulation-Guided Reinforcement Learning

Green Code or Costly Code? Evaluating the Energy Costs of Iterative LLM-Based Code Optimizations