MODEL LIBRARY
Find the right AI model for your next creation
Turn a prompt into an image, refine a reference, or bring a scene to life. Choose a model for the way you want to create.
Image models
14 modelsChoose an image model for your next creation or edit. Each page includes its own workbench and model details.
- FeaturedGPT Image
GPT Image 2.5 Flare
Create everyday visuals or edit reference images, with transparent background options.
Start creating -
FeaturedGrok Grok Imagine Image 2.0
Generate images from a prompt or guide a new composition with reference images.
Start creating
-
GPT Image 2.5 Sunburst
Create detail-focused images and refine existing visuals with reference-based editing. · 1K / 2K / 4K
-
Seedream 5 Lite
Combine image references with written instructions to create or edit a visual. · 2K / 3K / 4K
-
Seedream 5 Pro
Combine image references with written instructions to create or edit a visual. · 1K / 2K
-
Nano Banana 2
Combine image references with written instructions to create or edit a visual. · 1K / 2K / 4K
-
Nano Banana 2 Lite
Start with a text prompt or upload an image and describe the changes you want.
-
Nano Banana Pro
Combine image references with written instructions to create or edit a visual. · 1K / 2K / 4K
-
Nano Banana
Start with a text prompt or upload an image and describe the changes you want.
-
GPT Image 2
Start with a text prompt or upload an image and describe the changes you want. · 1K / 2K / 4K
-
GPT-4o Image
Start with a text prompt or upload an image and describe the changes you want.
-
Flux Kontext
Describe changes to a subject, scene or style using your existing image as context.
-
Seedream 4.5
Combine image references with written instructions to create or edit a visual. · 2K / 4K
-
Seedream V4
Start with a text prompt or upload an image and describe the changes you want. · 1K / 2K / 4K
Video models
10 modelsExplore video models and their generation settings. Open a model page to choose inputs, duration and output options.
- FeaturedSeedance
Seedance 2.5
Guide video creation with text, images, video clips and audio references.
Start creating - FeaturedWan
Wan 3.0 Video Prime
Create a video from text, first and last frames, or reference materials.
Start creating
-
Grok Video 1.5 Preview
Describe a moving scene or animate a still image with your motion instructions. · 480p / 720p
-
Seedance 2 Mini
Guide video creation with text, images, video clips and audio references. · 480p / 720p
-
MiniMax H3
Create video from text, images or mixed references, including video and audio. · 768p / 2K
-
Wan 3.0 Video
Create a video from text, first and last frames, or reference materials. · 480p / 720p / 1080p
-
Sora 2
Describe a moving scene or animate a still image with your motion instructions.
-
Veo 3.1
Describe a moving scene or animate a still image with your motion instructions.
-
Seedance v1 Lite
Describe a moving scene or animate a still image with your motion instructions.
-
MindVideo AI
Describe a moving scene or animate a still image with your motion instructions.
No matching models
Try a model name, a shorter keyword, or a different category.
Choosing and using models
Answers about inputs, output settings and credit costs.
Should I choose an image model or a video model?
Choose an image model for a poster, product visual, illustration or still image edit. Choose a video model when you need motion. You can first create an image, then use it as the starting image in a compatible video model.
How do Flare and Sunburst differ?
Flare is a practical starting point for everyday generation and editing. Sunburst focuses on intricate detail and may take longer. Both offer 1K, 2K and 4K output and reference image editing here.
Can I upload several reference images?
Yes, on models that accept multiple references. GPT Image 2.5 accepts up to 16; Grok Imagine Image 2.0 accepts up to 5. Choose a model and check its upload slots before preparing your references.
Which models offer transparent backgrounds?
GPT Image 2.5 Flare and Sunburst include a transparent background setting. Select it in the workbench and describe the subject clearly. Keep a PNG or WebP version when you need to preserve transparency.
How should I choose resolution and aspect ratio?
Match the ratio to the placement: landscape for a wide frame, portrait for a mobile feed, or square for a product tile. Start at a lower resolution to check composition, then choose a supported higher resolution for the final image.
How is image-to-video different from reference-to-video?
Image-to-video usually starts from a still image that you want to animate. Reference mode uses images, clips or audio to guide appearance, movement or rhythm. Seedance 2.5, Seedance 2 Mini, Wan 3.0 and MiniMax H3 offer reference workflows with different upload limits.
Can I set the beginning and end of a video?
Models with first and last frame controls let you upload a starting image and an ending image. Select frame mode, keep the scene and subject compatible, and describe how the camera or subject should move between them.
How are generation credits calculated?
Costs depend on the model and selected settings. Some images are priced by resolution; video costs depend on duration and resolution. Reference video duration can also affect the total, and MiniMax H3 charges for images beyond the first five. Review the estimate before submitting.
How do I compare models without spending too many credits?
Use the same prompt and references, choose the smallest suitable output, and compare composition, text and subject consistency. For video, try a short clip before increasing duration. Adjust one setting at a time.
Where can I find the result or improve it?
Open generation history to view and download completed work. For revisions, use the result as a new reference and specify what to keep and what to change. For video, give separate instructions for subject motion and camera movement.