Choose the AI model that builds your app
In shortBeside Build it, the model button reads Auto — MonstarX then picks the right model and effort for every request and shows its choice on each reply. Pick a model yourself when you want one model every time. Premium models use more credits and come with the Plus, Pro and Ultra plans.
An AI model is the “brain” that writes your app’s code and answers you. MonstarX offers many, from quick and inexpensive to the most capable. You do not have to choose: Auto picks for you on every request. This page explains what Auto does, when to pick a model yourself, and how the choice affects your usage.
What does Auto do?
Section titled “What does Auto do?”Auto looks at each request and picks the model and the effort that suit it: a quick, low-cost model for a question or a small tweak, a strong building model for a new feature or a whole app, and a model known for careful fixing when something is broken. Every reply shows what was picked, like Auto · Claude Sonnet 5 · Medium.
If a build falls behind — checks keep failing, or it is running out of steps before its checklist is done — Auto can step up to a stronger model for the rest of that request, and the reply says so. For most people, Auto is the best choice.
How do I pick a model myself?
Section titled “How do I pick a model myself?”-
In the prompt box on Projects, click the model button (it reads Auto until you change it).
-
Pick a model from the list. The list opens with Auto at the top, then the models in groups.
-
If the model offers effort levels, a second button appears beside it. Pick how long the model thinks (see below).
-
Describe your app and press Build it. The new project builds with the model you picked.
A model picked this way belongs to that one project. The next new project starts on Auto again. To change the model of an existing project, use the model button beside the chat box in its workspace, or Project settings.
What do the groups mean?
Section titled “What do the groups mean?”| Group | What it says | When it fits |
|---|---|---|
| Fast ($) | Quickest replies | Small changes and questions when speed matters most. |
| Lower cost ($) | Fewest credits per request | Getting a lot done on a small plan. |
| Standard ($$) | More credits per request | Strong, dependable building for most apps. |
| Premium ($$$) | Most credits per request | The most capable models, for the hardest work. |
Each model shows its maker and a hint beside its name — 1M context (it can read very long material at once), coding, deeper reasoning. A No images tag means the model reads text only, so pictures cannot be sent to it. The groups are set by what each model costs to run, so they stay accurate as models change.
What is the effort level?
Section titled “What is the effort level?”Effort is how long a model thinks before it answers. It appears beside a model you picked, when that model offers levels:
| Level | Trade-off |
|---|---|
| Low | fastest, fewest credits |
| Medium | balanced |
| High | thinks longer, more credits |
| Extra high | thinks much longer, more credits |
| Max | deepest thinking, most credits |
The level marked (default) is what the model uses if you do not choose. Models offer different levels, and a few offer none. With Auto, effort is chosen for you on each request.
When should I pick a model myself?
Section titled “When should I pick a model myself?”- You have a favourite model and want every request to use it.
- You want to keep usage low — pick a Lower cost or Fast model.
- You are working on something unusually hard and want a Premium model on every step.
A model you picked stays yours: MonstarX never swaps it for a stronger one. (Only if it is unavailable or stops answering does the request carry on with another model, and the reply says so.) If a build on it falls well behind its checklist, the reply says so and suggests Auto, which moves that kind of build to a stronger model.
How do premium models count toward my usage?
Section titled “How do premium models count toward my usage?”Every request uses some of your plan’s usage, and bigger models use more. Premium models also draw on a weekly premium allowance:
- Free — premium models are locked; the group shows Plus plan and up with See plans.
- Plus — a small weekly allowance to try them; the group shows how much of it you have used.
- Pro and Ultra — premium models can use all of your usage; the group shows Included in Pro or Included in Ultra.
When the weekly premium allowance runs out, pick a model from Lower cost or Standard until it resets. See Usage and limits and Plans and pricing.
Frequently asked questions
Section titled “Frequently asked questions”Which model does Auto use?
It depends on the request, and it changes as better models arrive. Each reply shows the model and effort that did the work, for example Auto · Claude Sonnet 5 · Medium.
Can I switch models halfway through a project?
Yes. Change the model beside the chat box or in Project settings. The next request uses it; earlier work is unchanged.
Why does a model have a lock beside it?
It is a premium model that your plan cannot use right now: the Free plan has no premium models, and on Plus the weekly premium allowance may be used up. Upgrade, wait for the weekly reset, or pick a model from another group.
Does a higher effort always give a better app?
Not always. Higher effort helps with tricky logic and hard bugs, but it is slower and uses more credits. For most building, the default level is the best balance.
What does No images mean?
The model reads text only. Pictures you attach cannot be sent to it — pick another model (or Auto) when your request includes a mockup or screenshot.