This document provides an overview of the AI models available in Builder.io’s platform, highlighting their strengths, relative costs, and where to find current pricing.
For specific information on how external pricing converts to credits, see Agent credits.
An AI model is the engine that powers Builder.io's generative AI operations. Model choice drives quality, speed, and cost for every AI operation across the platform.
When choosing between models, two factors matter most:
- Quality: higher-quality models produce more accurate, reliable reasoning and output.
- Cost: each model has a relative cost, which is converted into Agent credits.
There are several AI models available.
The Auto setting automatically selects the best available model for your task. Free users always use Auto, and it is the default for self-serve plans.
This flexibility means Builder.io's team can continuously improve efficiency while keeping quality high, so you get more value out of your Agent credits.
Claude Sonnet 4 (default)
Claude Sonnet 4 delivers strong reasoning and reliable code generation, making it well-suited for complex, multi-step development workflows. For more information, see Anthropic's pricing.
GPT-5
GPT-5 provides a balance of cost and quality. It is best suited for general reasoning and long-form tasks across domains.
GPT-5 Mini
GPT-5 Mini is a smaller, faster variant designed for lightweight tasks where cost efficiency is critical. It is well-suited for quick iterations and high-volume requests. For more information, see OpenAI's pricing.
There are two places where you can select the AI model you want to use: on the main Projects page or within a Project that's already running.
To select a particular AI model:
- Go to Projects.
- In the main Projects input, click the AI model dropdown and select the model you'd like to use.
The video below shows where the AI model selector is in Projects:
For an introduction to using Projects, visit Projects overview.
To select a particular AI model:
- Go to Projects.
- With your project running, click the AI model dropdown and select the model you'd like to use.
The video below shows where the AI model selector is when you are in a project that's running:
For an introduction to using Projects, visit Projects overview.
All external model costs are converted to agent credits using Builder.io's transparent formula. For details on how token costs convert into credits, see Agent credits.
Performance or optimization questions: reach out through the in-app support chat or email support@builder.io.
Enterprise customers: contact your account team for guidance on model selection and usage planning.
By default, the system selects Claude Sonnet 4. If you manually choose another model from the dropdown, that choice will be reflected during your session.
If Claude Sonnet 4 is temporarily unavailable, the system will fall back to another high-quality model to keep you unblocked.
Yes. Some models may be restricted by plan or availability. Enterprise customers can request access to specific models through their account team.
Yes. Most models support multiple languages, but quality can vary. For the strongest multilingual performance, consider models like Claude Sonnet 4, GPT-5, or Z-AI GLM 4.5.
Credit conversions and examples are covered in detail in Agent Credits.
Auto is the default option for Free and self-serve users. It automatically chooses the most cost-effective model for your task.