The best AI model for diy and home repairs
A shopping list and a safety stop, not a license.
Short answer
ChatGPT Free ($0) and GPT-5.6 Luna are the subscription fit. On the API, GPT-6 Luna at $0.10 in · $0.50 out per million tokens The rate is for 1,000,000 tokens, not for one message. A short email is a tiny fraction of that. A long contract with a long answer can use a real slice of it. Glossary is the cheapest sensible pick. This illustration A chunk of text the bill is counted in. A token is often a short word or part of a word. One million tokens is roughly 750,000 English words. That ratio is a rule of thumb, not a count of your file. Glossary is 500 input tokens and 700 output tokens, about $0.0004.
For a wobbly shelf, a running toilet, or a picture you want hung straight, a model can turn your description into steps and a parts list. Gas, the electrical panel, a roof, anything structural, and anything that can fall on a person are stop signs. Hire a licensed person. A larger model is not that person.
Subscription
ChatGPT Free and GPT-5.6 Luna are enough for a photo and a parts list. The help center says image uploads are on Free and paid plans. A paid plan does not license you to do electrical or gas work.
ChatGPT plans · Check a plan you already pay for
API
The typical case is a photo of the problem plus a short question. GPT-6 Luna is the cheapest reviewed model on this feed that lists both file and image input. DeepSeek V4 Flash is cheaper and text-only, so it is the wrong default when the input is a picture. The illustration is a short prompt and a parts list. Claude Haiku 5.5 lists the same $0.10 in · $0.50 out on this feed. Luna is the named pick so the page has one default.
GPT-6 Luna lists a context window How much text the model can hold in one request: your prompt, the files, and the reply together. If the document is bigger than the window, it does not fit in one go. Glossary of 1,050,000 tokens. A short everyday paste uses a sliver of that.
| Lane Quality, balanced, and cheapest sensible are editorial lanes, not a lab’s product names. Quality is the reviewed pick for hard work. Cheapest sensible is the reviewed pick when a miss is easy to catch. Glossary | Model | Rate What you pay for tokens you send. $2 in means $2 per million tokens of your prompt, files, and images. A rough English guide is about 750,000 words per million tokens. Your file can be shorter or longer. This is not a measurement of a specific document. Glossary | This illustration What you pay for tokens the model writes back. $10 out means $10 per million tokens of the reply. Long answers cost more than short ones, even when the question was small. Glossary |
|---|---|---|---|
| Cheapest sensible | GPT-6 Luna | $0.10 in · $0.50 out | $0.0004 |
| Step up | GPT-6.2 Sol | $2.00 in · $10.00 out | $0.0080 |
Input is what you send What you pay for tokens you send. $2 in means $2 per million tokens of your prompt, files, and images. A rough English guide is about 750,000 words per million tokens. Your file can be shorter or longer. This is not a measurement of a specific document. Glossary. Output is what the model writes back What you pay for tokens the model writes back. $10 out means $10 per million tokens of the reply. Long answers cost more than short ones, even when the question was small. Glossary. $0.10 in and $0.50 out are per million tokens, not per message.
Step up when
- If you can describe the problem in words, DeepSeek V4 Flash is the cheaper text-only model. Luna is the default here because the usual input is a photo.
- Luna misreads the photo. Say what it got wrong. If it still misreads, try Sol, which also lists image input, and then decide whether this is a job for a person.
- The repair needs a manual. Upload the manual’s relevant pages rather than asking the model to remember your appliance.
- Water is near electricity, you smell gas, or the damage is structural. Stop. Do not step up to a flagship. Step out of the chat.
The step-up model on this page is GPT-6.2 Sol, $2.00 in · $10.00 out per million tokens. The same illustration is about $0.0080.
Try these prompts
The wobbly shelf
The shelf in this photo pulls out of the drywall when I put books on it. What parts should I buy for a drywall anchor a renter can install? Tell me when to stop and call someone. Do not assume I own tools I did not name.
The running toilet
The toilet runs for a minute after a flush. List the usual checks in order, starting with the flapper. Tell me which step means I should close the water and call a plumber.
The stop sign
I smell gas near the stove. What should I do in the next five minutes? Do not give me a repair procedure.
Overkill
In the app you already use
Only features with a vendor page. Checked 20 Nov 2026.
- A photo in the chat — OpenAI documents image uploads on Free and paid plans. Claude’s upload article also lists JPEG, PNG, GIF, and WebP. A photo is not a diagnosis of the wiring behind the wall.
- Claude image upload — Claude’s help center lists those image types and a 500 MB chat-file limit. Haiku is on Claude Free if you want the light model in that app. The file article does not cover gas or electrical safety.
Questions
Can I use it for the breaker box?
No. Use it to write down symptoms for an electrician. Do not follow a procedure on the panel from a chat.
Why is the API pick not the cheapest text model?
Because the usual input is a photo. The cheaper text model does not list image input. If you can describe the problem in words, the text model is the cheaper illustration, and this page says so in the step-up section the other way around.