In short
Ask for what you need in one clear request, keep answers short, start a new chat for a new topic, and go easy on image and video generation. A typical text question uses roughly a third of a watt-hour on published estimates, but long inputs, repeated regeneration and media generation can multiply that many times over.
Steps
- Check whether a search, calculator or map would answer the question faster.
- Write one clear request that says what you want and how long the answer should be.
- Refine the answer you have instead of pressing regenerate.
- Start a fresh chat for a new topic rather than continuing a very long thread.
- Use the lighter or faster model option for simple questions.
- Generate images and video only when you actually need them.
You can cut most of the energy of your ChatGPT use by changing how much it reads and writes, and how often it re-runs. One ordinary question is small: OpenAI's chief executive has said an average query uses about 0.34 Wh, and Epoch AI independently estimated about 0.3 Wh. The habits that multiply that figure are what matter.
What uses the most energy?
Long inputs, long outputs, repeated attempts and media. Epoch AI estimates that a query with a 10,000-token input (a long pasted document) rises to about 2.5 Wh, and one with a 100,000-token input to almost 40 Wh. In tests by Luccioni, Jernite and Strubell, generating images used on average about 2.9 kWh per 1,000 images, against 0.002 kWh for classifying 1,000 pieces of text.
How do I use it more lightly?
- Ask once, clearly. Say what you want, who it is for and how long the answer should be.
- Ask for short answers. UNESCO and University College London found that shorter prompts and responses could cut energy use by more than half.
- Paste only what is needed, such as the relevant section rather than the whole document.
- Pick the lighter model for simple questions, and save "thinking" modes for hard ones (see when to use reasoning modes).
What makes the biggest difference?
Avoiding loops and media. Pressing regenerate ten times, or making dozens of image variations, costs far more than any number of polite words. Per-query figures are estimates and vary by model and provider, so treat them as a rough guide, not a precise count.
For more everyday habits, read use AI efficiently and our plain-English guide for everyone.
Sources
- The gentle singularity, Sam Altman (OpenAI) (2025)
- How much energy does ChatGPT use?, Epoch AI (2025)
- Smarter, smaller, stronger: resource-efficient generative AI and the future of digital transformation, UNESCO and University College London (2025)
- Power Hungry Processing: Watts Driving the Cost of AI Deployment?, Luccioni, Jernite and Strubell, ACM FAccT (2024)
Related
How do I write prompts that use less energy?
How to write AI prompts that need less computation: be specific, set a length, paste only what matters and get it right first time instead of retrying.
How do I stop regenerating AI answers over and over?
Why repeatedly pressing regenerate wastes energy and time, and how to get a usable AI answer with one good prompt and a short follow-up instead.
How do I stop long AI chats wasting energy?
Why very long AI chat threads cost more with every message, and simple habits to keep them lean: new chats, short recaps and pasting only what matters.
Use AI efficiently: everyday habits
Everyday habits that cut the energy of using AI: clear prompts, fewer regeneration loops, care with images and video, caching and turning off defaults.