Scoring
Four numbers, one honest opinion.
Every builder gets the same brief: a five page site for a small studio, with a blog, a contact form and a real domain at the end. We score what comes out, not what the marketing page promises.
Craft
30%How good the finished site actually looks and feels, judged on a real build rather than a demo.
Speed
25%How quickly a blank prompt becomes something you would be willing to show a customer.
Control
25%How far you can steer the result: structure, styling, data, code and hosting.
Value
20%What you get before you pay, and whether the paid tier stays fair as the site grows.
Nothing on this page is sponsored. Nobody pays to be listed, nobody pays to move up, and every link is a plain link. If a tool improves, its score moves. If it stops being maintained, it leaves the index.
The weighting is not the same on every index.
Web, mobile and tools share one weighting. The model index does not, because you are not buying a workflow from a model, you are buying the answer.
| Index | Craft | Speed | Control | Value |
|---|---|---|---|---|
| Web | 30% | 25% | 25% | 20% |
| Mobile | 30% | 25% | 25% | 20% |
| Tools | 30% | 25% | 25% | 20% |
| LLMs | 55% | 10% | 15% | 20% |
A wrong answer is worth nothing however fast or cheap it arrived, so craft carries most of the weight on the model index. Price and openness are still 35% between them, which is why an open model can finish above a closed one it narrowly trails — they just cannot carry a weak model over a strong one.
Questions
What people ask about the ranking.
How is each tool scored?
Every entry gets the same brief and the same four scores out of 100: craft, speed, control and value. The headline score is those four, weighted. Rank is that weighted score and nothing else, so nothing can be promoted by hand.
Why is the model index weighted differently?
Web, mobile and tools weight craft 30, speed 25, control 25 and value 20. Models are weighted craft 55, speed 10, control 15 and value 20, because a model is bought for the answer rather than the workflow, and a wrong answer is worth nothing however fast it arrived.
Why is the publisher of this index ranked in it?
BestAIBuilder is published by Modulify AI, so the entry at the top of the web index is our own product. It is scored on the same rubric as everything else and gets no bonus for being ours. The disclosure sits on the entry itself, not in a footnote.
What does a provisional score mean?
A provisional entry is scored from the product’s public capability and pricing rather than from the hands-on rebuild the fully reviewed entries went through. Treat it as a placing rather than a verdict. Every provisional entry is labelled as one.
How often does the ranking change?
It is reviewed weekly, and every entry carries the date it was last edited. Editing a score in the CMS re-ranks that index on the next page load, so the ranking and the date always agree.
Is any placement sponsored or affiliate?
No. Nobody pays to be listed, nobody pays to move up, and every outbound link is a plain link. If a tool improves, its score moves. If it stops being maintained, it leaves the index.
Can I quote these scores?
Yes. Quote any score, ranking or comparison on the site, with a link back to the page it came from. Scores move week to week, so the date on the page is the date the figure was true.