# Study measures energy use of 86 open-weight LLMs on benchmarks and hardware

> Researchers found many LLM setups waste 2.6–4.8× more energy than the most efficient ones, with some using 27× more.

Oossa · 2026-10-14 · https://oossa.com/en/study-measures-energy-use-of-86-open-weight-llms-on-benchmarks-and-hardware

A team led by Zeyu Yang published a paper on Oct 14, 2026 that measured how much electricity 86 open-weight large language models (LLMs) consume while running seven benchmarks on six hardware platforms. The study shows most configurations use far more power than needed – a median of 2.6 to 4.8 times more, and up to 27 times more in extreme cases. Adding reasoning steps raises energy use by about 3.7×, while using low‑precision FP8 arithmetic cuts energy by roughly 25% with little loss in score.

## The facts

- 86 model checkpoints evaluated
- median energy waste 2.6–4.8×; max 27×

## Why it matters

Developers can cut operating costs and carbon footprints by choosing more energy‑efficient LLM setups.

## Sources & references

1. [Measuring the Inference Energy Efficiency of Open-Weight LLMs across Benchmarks, Reasoning Modes, and Hardware](https://arxiv.org/abs/2610.14389) – arXiv, 2026-10-14

Last updated: 2026-10-14