Image Models
Model Lab's Image Mode lets you explore and compare image generation models in one place. Switch between Text-to-Image and Image-to-Image, and adjust parameters for each model to find the best fit for your use case.
Test Image Models
Follow these steps to generate images and compare outputs:
- Select a mode: Select
Text-to-ImageorImage-to-Imageto get started. - Select models: Click Add Models to choose one or more models. Available categories include:
- Model Studio models: Requires API key configuration. See Configure API Keys.
- Your deployed models: All models you have deployed on Smart Studio.
💡 Tip: Which model should I choose?
- Qwen Image Models: Best for rendering complex text inside images and general image editing.
- Wan models (e.g., Wan2.6): Best for precise style control, understanding long prompts, and advanced multimodal generation.
To learn about other models, see How to choose models
-
Configure generation: Enter your text prompt. In Image-to-Image mode, upload a reference image or paste an image URL. These core inputs apply to all selected models. Optionally, select a model in the left panel to adjust its settings. Available options vary by mode. See Example Parameters for details.
-
Generate and compare: Click Generate to start the process. Once complete, review the outputs to compare visual quality, style adherence, and generation speed across your selected models.
-
View code (Optional): Click View Code to instantly generate an API request script containing your exact prompts and parameter configurations.
Generation costs vary depending on your selected model and the number of images generated (the N parameter). Review Model pricing before running large-scale tests.
- Standard generation: Cost = Unit Price ×
N
This pricing applies to Alibaba Cloud Model Studio models only. Self-deployed models are not included.
Example Parameters
Adjust these settings to generate new images from text prompts.
| Parameter | Description |
|---|---|
| N | Sets the number of images to generate per request. Supported values depend on the model Tip: Set this to 1 during initial testing to minimize API costs. |
| Seed | Sets the random seed for image generation (0 to 2147483647). Leave blank to use a random seed. |
| Size | Sets the output resolution (width × height). Supported values vary by model. |
| Watermark | Adds a visible watermark to the bottom-right corner of the generated image. |
| Prompt Extend | Automatically refines and expands your positive prompt to improve generation quality. This feature does not affect the negative prompt. |
| Negative Prompt | Defines specific elements, styles, or subjects to exclude from the generated image. Supports up to 500 characters in English and Chinese (excess characters are truncated). Example: low resolution, low quality, deformed limbs, chaotic composition. |
The parameters listed here are examples for reference only. Available parameters vary by model. For the full parameter list, refer to the API documentation for your specific model. For example, see the Model Studio API Reference.