Salvato in:
Dettagli Bibliografici
Autori principali: Pathania, Priyavanshi, Mehra, Rohit, Sharma, Vibhu Saujanya, Kaulgud, Vikrant, Nevels, Tiffani, Podder, Sanjay, Burden, Adam P.
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2603.02949
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866914365399105536
author Pathania, Priyavanshi
Mehra, Rohit
Sharma, Vibhu Saujanya
Kaulgud, Vikrant
Nevels, Tiffani
Podder, Sanjay
Burden, Adam P.
author_facet Pathania, Priyavanshi
Mehra, Rohit
Sharma, Vibhu Saujanya
Kaulgud, Vikrant
Nevels, Tiffani
Podder, Sanjay
Burden, Adam P.
contents Large Language Models are rapidly gaining traction in software engineering, yet their growing carbon footprint raises pressing sustainability concerns. While training emissions are substantial, inference quickly surpasses them due to the sheer volume of prompts processed. This shift underscores the urgent need for accurate, prompt-level carbon measurement during inference to enable informed, sustainability-focused decision-making. To address the limitations of existing approaches, in this paper, we outline the guiding principles for a novel reference framework for LLM inference carbon estimation that can guide the design of future tools and provide a systematic foundation for advancing sustainability research in this domain. We also introduce SEAL, an early embodiment of these principles that leverages a multi-benchmark-driven approach for per-prompt carbon estimation. Its initial validation shows promising results, positioning SEAL as a foundation for standardized sustainability assessment across the LLM ecosystem.
format Preprint
id arxiv_https___arxiv_org_abs_2603_02949
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SEALing the Gap: A Reference Framework for LLM Inference Carbon Estimation via Multi-Benchmark Driven Embodiment
Pathania, Priyavanshi
Mehra, Rohit
Sharma, Vibhu Saujanya
Kaulgud, Vikrant
Nevels, Tiffani
Podder, Sanjay
Burden, Adam P.
Software Engineering
Artificial Intelligence
Large Language Models are rapidly gaining traction in software engineering, yet their growing carbon footprint raises pressing sustainability concerns. While training emissions are substantial, inference quickly surpasses them due to the sheer volume of prompts processed. This shift underscores the urgent need for accurate, prompt-level carbon measurement during inference to enable informed, sustainability-focused decision-making. To address the limitations of existing approaches, in this paper, we outline the guiding principles for a novel reference framework for LLM inference carbon estimation that can guide the design of future tools and provide a systematic foundation for advancing sustainability research in this domain. We also introduce SEAL, an early embodiment of these principles that leverages a multi-benchmark-driven approach for per-prompt carbon estimation. Its initial validation shows promising results, positioning SEAL as a foundation for standardized sustainability assessment across the LLM ecosystem.
title SEALing the Gap: A Reference Framework for LLM Inference Carbon Estimation via Multi-Benchmark Driven Embodiment
topic Software Engineering
Artificial Intelligence
url https://arxiv.org/abs/2603.02949