In short
Start a new chat when you change topic, and when a long conversation is still useful, ask for a short recap and carry on from that in a fresh chat. Chatbots typically reread the conversation so far to answer each new message, so very long threads make every reply more expensive than it needs to be.
Steps
- Start a fresh chat for each new topic or task.
- Ask for a five-line summary when a long chat has reached a useful point.
- Paste that summary into a new chat and continue from there.
- Remove large pasted documents from the conversation once you have what you need from them.
- Avoid keeping one "everything" chat open for weeks.
A long chat gets more expensive with every message. To answer you, a chatbot typically has to process the conversation so far, not just your latest line. After dozens of turns and a pasted report or two, each new reply can be reading many thousands of tokens of history.
How much does input length matter?
A lot. Epoch AI estimates a typical ChatGPT query at about 0.3 Wh, rising to about 2.5 Wh with a 10,000-token input and almost 40 Wh with a 100,000-token input. A very long thread with pasted documents can push every reply towards the higher figures. Some services cache or trim history behind the scenes, so the real cost varies and is rarely published.
How do I keep chats lean?
- One topic, one chat. It also makes answers more focused, because the model is not distracted by old context.
- Recap and restart. Ask "Summarise what we've agreed in five lines", then open a new chat with that summary.
- Paste extracts, not whole files. Once you have pulled out what you need, you do not need the original in the thread.
- Use projects or saved instructions where your tool offers them for preferences you would otherwise repeat.
Common mistakes
- Keeping one giant chat as a notebook. Save notes elsewhere and bring in only what is relevant.
- Re-pasting the same document several times in one thread.
- Assuming memory is free. Anything the model has to take into account still has to be processed.
If you build chat tools, the same idea applies in code: see stop your chatbot resending the whole conversation. More tips in use AI efficiently.
Sources
- How much energy does ChatGPT use?, Epoch AI (2025)
- What are tokens and how to count them?, OpenAI Help Center (2025)
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 my chatbot resending the whole conversation?
How to keep chatbot context small: cap past turns, summarise older messages, retrieve what's relevant and cache fixed prompts, so input stops growing.
How do I reduce the energy my ChatGPT use consumes?
Practical ways to cut the energy of using ChatGPT and similar chatbots: shorter exchanges, fewer regenerations, lighter models and less image generation.
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.