AI Energy Calculator

See how much energy, carbon and water your team's AI use adds up to, and how much the model you pick changes it.

1What do you use AI for?
2How many people use it?
3Prompts per person a day
Pick at least one use in step 1 to continue.
Answer the questions, then press See my footprint

How the calculator works

Energy per token

Each model has its own energy use per input and output token, from Greenpixie’s GPU benchmarks. It includes data centre overhead and a share of training.

Typical prompts per task

Research, documents, coding and agents send and receive very different amounts of text. Each task uses a typical prompt size for that kind of work, and your prompts per day count for each use you pick. Totals assume 250 working days a year.

Carbon and water

Energy is turned into carbon using the latest full month of grid data, averaged across cloud regions, plus a share of hardware manufacturing. Water covers data centre cooling and power generation.

The methodology has been independently verified in accordance with ISO 14064-3, prepared with reference to ISO 14064-1, and designed to be in line with the GHG Protocol Corporate Standard. Read the full AI methodology.

How we built the data

To get a reliable picture of AI energy use, we ran our own benchmarks across a wide range of models and workloads. The AI methodology docs set out every step in full.

Step 1 of 6

Benchmarked open models on current GPUs

32 open-weights LLMs, from 1.7 billion to 1 trillion parameters, run quantised (FP8, INT8, NVFP4) on two generations of current GPUs, the way hyperscalers deploy them.

AI energy per prompt: key figures

A typical chat prompt to a large frontier model such as GPT-6.1 Sol uses about 1.6 Wh of electricity, about 8 mL of water and about 0.7 g CO₂e. It sounds small, but it adds up fast across a team, and some models and tasks use many times more. Our guide to AI energy consumption explains what drives it.

One chat prompt
≈1.6 Wh
About ten minutes of a 10 W LED bulb
Snapshot from Greenpixie’s AI methodology, November 2026. Includes data centre overhead, a share of training and hardware manufacturing. The calculator above uses the latest figures.
One chat prompt toElectricityCarbonWater
GPT-6 Astra7.0 Wh3.3 g CO₂e34 mL
GPT-6.1 Sol1.6 Wh0.74 g CO₂e8.1 mL
GPT-6 Luna0.19 Wh0.08 g CO₂e0.95 mL
  • ~50×GPT-6 Astra uses about 50 times the energy of GPT-6 Luna for every output token.
  • 0.34 WhOpenAI's CEO has said an average ChatGPT query uses about 0.34 Wh, without saying how that was worked out. Our figure for a large frontier model is nearly five times that.
  • 500+AI models covered by Greenpixie’s dataset, across every major provider.

About Greenpixie

Greenpixie is an enterprise data company that helps organisations understand and cut the environmental impact of their cloud use. We combine cost, carbon, energy and water into one consistent method that teams can use across regions and providers. AI is the next frontier of cloud use, and we hold it to the same standard as the rest of the cloud estate.

  • Resource and instance-level data
  • One method across AWS, Azure and GCP
  • Includes data centre PUE and Scope 3 (embodied) emissions
  • Uses granular grid data for location and time
  • ISO 14064-3 verified
  • Delivered daily, built for operational use
Having been involved in the Sustainable AI Working Group and seen the methodology up close, it’s clear Greenpixie are setting the gold standard for how enterprises can measure the energy, carbon and water impact of AI workloads.
Mark BradleySenior Manager, Product, Flexera
The methodology strikes the right balance between scientific rigour and enterprise usability. It gives organisations a credible way to understand and optimise the sustainability impact of AI, including consumption of water, kWh and carbon.
Andy WestbrookDirector, Cloud Transformation & FinOps, NTT Data

Frequently asked questions

Quick answers on AI energy, water and how this calculator works.

How much energy does a ChatGPT prompt use?
A typical chat prompt to a large frontier model such as GPT-6.1 Sol uses around 1.6 Wh of electricity based on Greenpixie's AI methodology. That includes data centre overhead and a share of the energy used to train the model. A small model such as GPT-6 Luna uses under 0.2 Wh for the same prompt.
How much water does AI use?
Water is used to cool data centres and to generate the electricity they run on. In Greenpixie's data, the same 1.6 Wh chat prompt to GPT-6.1 Sol uses about 8 mL of water. Agentic workflows, which send far more tokens, can use hundreds of millilitres per task on the largest models. Our guide to AI water usage goes deeper.
Why does the model make such a big difference?
Bigger models run on more GPUs for longer to produce each token. In Greenpixie's data, GPT-6 Astra uses about 50 times the energy of GPT-6 Luna for every output token. Choosing the smallest model that does the job well is the biggest single way to cut AI energy use. See our sustainable AI guide for more ways to cut it.
How accurate is this calculator?
It gives a fair estimate for typical usage. The per-token figures come from Greenpixie's AI methodology, and each task type assumes a typical prompt and response size. Your real footprint depends on your actual prompts, the region and hour your requests run in, and the tools you use, which the Greenpixie app measures from your own usage data.

Beyond the estimate

Measure the real footprint of your AI

Every major model, your actual usage, and the region and hour each request runs in. One method across providers, SaaS tools and agents.