Select a model to power your results
AI generation models turn your prompts and reference content into images. Each model has its own characteristics and style. Some excel at rich, detailed artwork, while others are optimized for speed, realism, or editing precision.
The available models vary depending on the workflow, the type of image you want to create, and whether you use only a prompt or include reference content. The table below highlights the strengths, performance, and recommended applications for each model.
|
|
|
|
Stable Diffusion 3.5 Large |
High-quality text-to-image generation with strong detail, rich textures, and reliable handling of complex scenes, lighting, and compositions. Moderate speed. |
High-quality creative work such as concept art, marketing visuals, stylized illustrations, and detailed scenes where fidelity and artistic richness matter more than speed |
Stable Diffusion 3.5 Medium |
Faster and less resource-intensive than Stable Diffusion 3.5 Large, with lower fidelity and fewer fine details in complex scenes but still capable of producing versatile, good-quality images. Fast. |
Balanced results for simpler images or quick iterations |
|
|
Extremely fast generation optimized for producing many variations and exploring ideas quickly. Less effective for precise, detail-oriented edits. |
Fast drafts, quick ideation, and high-speed concept exploration for workflows where rapid iteration takes priority over detail or accuracy |
|
|
High-fidelity generation with strong prompt accuracy, excellent reference image interpretation, and advanced editing and refinement capabilities. Moderate speed. |
Production-quality images, particularly when you need detailed realism, precise edits, consistent characters or objects across versions, or transformations guided by reference images |
|
|
Focused on guided editing and transformations, with emphasis on keeping key elements in reference images. Good for short, concise prompts. Very fast. |
Lightweight editing, concepting, quick mockups, campaign assets, visual refinements, stylistic adjustments, and scenarios where preserving key aspects of an input image is important |
|
|
Supports 2K image generation and advanced text rendering. High prompt accuracy. Very strong reference image fidelity. Fast. |
High-fidelity, production-level visuals — campaigns, marketing assets, posters with text, professional infographics, complex edits |
|
|
High-fidelity image generation and editing — precise prompt adherence, consistent style, strong reference fidelity. Very fast. |
High-quality image generation and editing, creative projects, marketing visuals, brand-consistent content, detailed prompts, and preserving key aspect of an input image |
For detailed information about each model, visit the official model pages.
Choosing the right model helps you get the results you want more quickly and reliably — whether that’s a complex design, a rapid concept sketch, a detailed illustration, a realistic photo, or an image you plan to edit and refine over multiple iterations. Cost is also important: each model consumes a different number of credits per image. Understanding model capabilities and credit usage lets you balance creative control with your budget. For information about credit allocation and usage, see Manage AI credits.
To choose a model
•
Select a model from the
Model list box.
The cost per image appears below each model.
Credit costs per image may change. For the latest, up-to-date information, visit the FAQ.
Model availability is subject to change; models may be added or removed. For the most current information, visit the FAQ.
The selected model is maintained from one session to the next.
The cost per image, the total cost for the selected number of images, and your remaining credits appear next to the Images list box.