Choose a model
Select the right AI model for your image, video, vector, or other generation task.
Why model choice matters
Different AI models are trained for different purposes and produce different results. Choosing the right model for your task affects:
- Output quality and style — each model has distinct aesthetic characteristics
- Generation speed — some models are faster; others produce higher quality at slower speeds
- Supported inputs — some models accept text only; others accept an image reference or both
- Features — specific parameters (like aspect ratio, style presets, or motion intensity) are model-specific
How to choose a model
- Open the prompt input.
- Select a creation tool (Image, Video, Vector, etc.).
- Click the model selector to open the model list.
- Browse available models — descriptions and tags indicate their strengths.
- Select a model.
- Configure the model-specific parameters that appear.
- Generate.
Model categories
| Category | When to use |
|---|---|
| Image models | Photo-realistic images, illustrations, concept art, product shots |
| Vector models | Icons, logos, scalable shapes, SVG-style artwork |
| Video models | Animated clips from text descriptions or starting images |
| 3D models | Product scenes and 3D object renders |
Reading model descriptions
The model list shows names, provider badges, and capability tags. Look for tags like:
- Text-to-image / Image-to-image — whether the model accepts image input
- Fast / Quality — speed vs. quality trade-off indicators
- Style labels — photorealistic, illustration, anime, etc.
Tips
- If you are new to a task type, try the default model first — it is typically a balanced choice.
- For final production-quality output, test 2–3 models and compare results at your target aspect ratio.
- Some models have strong aesthetic preferences (e.g., a model tuned for photorealism may struggle with flat illustrations). Match the model to the style you want.
- Model availability may change as new models are added. See Available models for the current list.