# Scaling Laws for Energy Efficiency of Local LLMs

arXiv: https://arxiv.org/abs/2512.16531 · HTML: https://mlsanity.com/paper/2512.16531/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Visual token pruning and efficiency, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 18 of its 36 references that are in the map.
- First posted on arXiv 18 Dec 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Visual token pruning and efficiency](https://mlsanity.com/mllm/visual-token-pruning-and-efficiency/)

## Builds on (most cited first, 18 of 18)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Flamingo: a Visual Language Model for Few-Shot Learning](https://mlsanity.com/paper/2204.14198/) (2022)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities](https://mlsanity.com/paper/2507.06261/) (2025)
- [SimCSE: Simple Contrastive Learning of Sentence Embeddings](https://mlsanity.com/paper/2104.08821/) (2021)
- [Green AI](https://mlsanity.com/paper/1907.10597/) (2019)
- [A Survey of Model Compression and Acceleration for Deep Neural Networks](https://mlsanity.com/paper/1710.09282/) (2017)
- [Towards the Systematic Reporting of the Energy and Carbon Footprints of Machine Learning](https://mlsanity.com/paper/2002.05651/) (2020)
- [Communication-Efficient Edge AI: Algorithms and Systems](https://mlsanity.com/paper/2002.09668/) (2020)
- [Provable Scaling Laws for the Test-Time Compute of Large Language Models](https://mlsanity.com/paper/2411.19477/) (2024)
- [Scaling Inference-Efficient Language Models](https://mlsanity.com/paper/2501.18107/) (2025)
- [Intelligence per Watt: Measuring Intelligence Efficiency of Local AI](https://mlsanity.com/paper/2511.07885/) (2025)
- [Optimizing Inference Performance of Transformers on CPUs](https://mlsanity.com/paper/2102.06621/) (2021)
- [Sometimes Painful but Certainly Promising: Feasibility and Trade-offs of Language Model Inference at the Edge](https://mlsanity.com/paper/2503.09114/) (2025)
- [Inference Compute-Optimal Video Vision Language Models](https://mlsanity.com/paper/2505.18855/) (2025)
- [Accuracy and Consumption analysis from a compressed model by CompactifAI from Multiverse Computing](https://mlsanity.com/paper/2507.08836/) (2025)

## Built on by (most cited first, 4 of 4)

- [Selecting Language Models for Social Science: Start Small, Start Open, and Validate](https://mlsanity.com/paper/2601.10926/) (2026)
- [The Energy Consumption of Transformer Fine-Tuning: A Roofline-Inspired Scaling Model](https://mlsanity.com/paper/2606.23546/) (2026)
- [Retrieval-Augmented Large Language Models as Components of Cognitive Computing architecture for Regulatory Knowledge Management](https://mlsanity.com/paper/2607.24352/) (2026)
- [Activation-Aware Weight Tensorization: A Calibration-Time Preconditioner for Tensor-Network LLM Compression](https://mlsanity.com/paper/2610.10085/) (2026)

## Abstract (arXiv)

Deploying local large language models and vision-language models on edge devices requires balancing accuracy with constrained computational and energy budgets. Although graphics processors dominate modern artificial-intelligence deployment, most consumer hardware--including laptops, desktops, industrial controllers, and embedded systems--relies on central processing units. Despite this, the computational laws governing central-processing-unit-only inference for local language and vision-language workloads remain largely unexplored. We systematically benchmark large language and vision-language models on two representative central-processing-unit tiers widely used for local inference: a MacBook Pro M2, reflecting mainstream laptop-class deployment, and a Raspberry Pi 5, representing constrained, low-power embedded settings. Using a unified methodology based on continuous sampling of processor and memory usage together with area-under-curve integration, we characterize how computational load scales with input text length for language models and with image resolution for vision-language models. We uncover two empirical scaling laws: (1) computational cost for language-model inference scales approximately linearly with token length; and (2) vision-language models exhibit a preprocessing-driven "resolution knee", where compute remains constant above an internal resolution clamp and decreases sharply below it. Beyond these laws, we show that quantum-inspired compression reduces processor and memory usage by up to 71.9% and energy consumption by up to 62%, while preserving or improving semantic accuracy. These results provide a systematic quantification of multimodal central-processing-unit-only scaling for local language and vision-language workloads, and they identify model compression and input-resolution preprocessing as effective, low-cost levers for sustainable edge inference.