Three moments when switching models in the AI image generator pays off.
Most briefs work on the default: prompt, optional references, 1K free. Switch only when a result keeps missing the same thing after two rewrites.
Got one photo to rework, or several to merge? Pick a model with an image-to-image or edit mode. Text-only models such as GPT Image 4o and Z-Image skip reference uploads.
Need 4K? Nano Banana and Seedream offer higher resolutions on paid plans. Need a quick batch of drafts? Some models return several images per run.
This AI image generator holds fifteen model versions in one workspace, from Google's Nano Banana family to ByteDance's Seedream, OpenAI's GPT Image, Black Forest Labs' Flux, Alibaba's Qwen, Z-Image and xAI's Grok Imagine. It opens on Nano Banana 2. The version bar lists the others, and each one brings its own modes, sizes and settings, so the controls below the prompt change when you switch. Every model needs a prompt. Most also accept reference images through an image-to-image or edit mode. Most versions are open on the free tier, with higher resolutions or larger models on paid plans. Results go to My Creations.
Each tab is a distinct input route.
Every model can build an image from a prompt. Start on Nano Banana 2 and compare a second model only when the first misses.
One Prompt Box
The same prompt carries across model switches.
Fifteen Versions
Seven model families in one list.
Retry on Another Model
Keep the wording, change the engine.
Six capabilities across the model list.
Google, ByteDance, OpenAI, Black Forest Labs, Alibaba and more.
Opens on the site's main model.
Image-to-image on most models.
On selected models and plans.
Most versions run without a paid plan.
Results from every model in My Creations.
Three groups, two workflows each.
Upload a photo of the empty room, pick an image-to-image model and describe the furniture style. Label staged photos as virtual staging in listings.
Use a wide format and describe the house front at golden hour, then add your text in a design tool. Keep the image honest to the property.
Photograph the plate on your phone, then use image-to-image to clean the background and light it like a menu shot.
Describe the theme and colours, generate a few options with a text-to-image model, and print the one that fits your shopfront.
Generate illustrations rather than photos when the subject is sensitive, so no real person is depicted. A friendly style works well for appeals.
Write one prompt per article and keep the same style words across them, so the issue looks consistent without a photo budget.
Three actions on this page.
Describe the image; add reference photos if the model supports them.
Stay on Nano Banana 2 or pick another version, then set its options.
Download from My Creations or reuse the result as a reference.
Choosing models, inputs and limits.
It is the site's main model, open on the free tier, and handles most everyday briefs with a prompt and optional references. Switch when a result keeps missing the same thing.
Most versions, including Nano Banana 2, GPT Image, Seedream 4.5, Flux 2 and Qwen. Nano Banana Pro and Seedream 5.0 Pro need a paid plan, and some higher resolutions do too. New accounts get 10 credits once, valid for 30 days.
The Nano Banana, Seedream, GPT Image 2 and 1.5, Qwen, Flux 2 and Grok Imagine versions. GPT Image 4o and Z-Image are text-only.
Up to 4K on Nano Banana and selected Seedream versions, on paid plans. Other models cap lower; the settings panel shows each model's options.
Each model has its own controls, so the panel only shows what the selected version supports, such as resolution, aspect ratio or batch size.
Commercial use follows our Terms of Service and your plan. You are responsible for rights in any photos you upload and for checking the result before publishing.
Dedicated pages for the most-used versions.
An AI image generator built around Nano Banana 2, with Seedream, GPT Image, Flux, Qwen, Z-Image and Grok Imagine one click away when a brief calls for a different input, size or look.
0 / 20000
