AI image models, compared
Omni Banana puts 31 AI image models in one studio, in your browser. Here is what each one can do. Open any of them to create from a description or edit your photos.
All models side by side
| Model | Made by | Create from a description | Edit your photos | Resolution |
|---|---|---|---|---|
| Nano Banana | Yes | up to 6 photos | Standard | |
| Nano Banana 2.1 | Yes | up to 14 photos | 1K, 2K, 4K | |
| Nano Banana 2 | Yes | up to 14 photos | 1K, 2K, 4K | |
| Nano Banana 2 Lite | Yes | up to 10 photos | Standard | |
| Nano Banana Pro | Yes | up to 14 photos | 1K, 2K, 4K | |
| GPT Image 1.5 | OpenAI | Yes | up to 4 photos | Standard |
| GPT Image 2 | OpenAI | Yes | up to 4 photos | Standard |
| GPT Image 2.5 | OpenAI | Yes | up to 4 photos | up to 3840x2160 |
| Grok Imagine | xAI | Yes | 1 photo | Standard |
| Grok Imagine Quality | xAI | Yes | 1 photo | 1K, 2K |
| Grok Imagine 2 | xAI | Yes | 1 photo | 1K, 2K |
| Seedream V4 | ByteDance | Yes | up to 10 photos | 1K, 2K, 4K |
| Seedream V4.5 | ByteDance | Yes | up to 14 photos | 2K, 4K |
| Seedream V5 | ByteDance | Yes | up to 14 photos | 2K, 3K |
| Seedream V5 Flash | ByteDance | Yes | up to 10 photos | Standard |
| Qwen Edit Plus | Alibaba (Qwen) | No | up to 6 photos | Standard |
| Qwen Image 2 | Alibaba (Qwen) | Yes | 1 photo | Standard |
| Qwen Image 2 Pro | Alibaba (Qwen) | Yes | 1 photo | Standard |
| Qwen Image 2.1 | Alibaba (Qwen) | Yes | up to 10 photos | Standard |
| Qwen Image 2.1 Pro | Alibaba (Qwen) | Yes | up to 10 photos | Standard |
| Ideogram V4 Turbo | Ideogram | Yes | No | up to 2048x2048 |
| Ideogram V4 Balanced | Ideogram | Yes | No | up to 2048x2048 |
| Ideogram V4 Quality | Ideogram | Yes | No | up to 2048x2048 |
| Recraft V3 | Recraft | Yes | No | Standard |
| Recraft V4 | Recraft | Yes | No | Standard |
| Recraft V4.1 | Recraft | Yes | No | Standard |
| Flux Kontext | Black Forest Labs | No | 1 photo | Standard |
| Flux Kontext Pro | Black Forest Labs | Yes | 1 photo | Standard |
| FLUX 3 | Black Forest Labs | Yes | up to 10 photos | 1K, 2K, 4K |
| Z Image Turbo | Pruna AI | Yes | No | Standard |
| P Image | Pruna AI | Yes | up to 4 photos | Standard |
The model families
Nano Banana, Nano Banana 2.1, Nano Banana 2, Nano Banana 2 Lite, Nano Banana Pro
Nano Banana is the name of Google’s Gemini image models. They are known for editing photos while keeping people and details consistent, and for combining several photos into one image.
GPT Image 1.5, GPT Image 2, GPT Image 2.5
GPT Image is OpenAI’s family of image models, the ones behind image creation in ChatGPT. They follow long, detailed prompts closely and handle text inside images well.
Grok Imagine, Grok Imagine Quality, Grok Imagine 2
Grok Imagine is xAI’s image model, the one used in Grok.
Seedream V4, Seedream V4.5, Seedream V5, Seedream V5 Flash
Seedream is ByteDance’s family of image models, made for high-resolution images and for editing with several reference photos.
Qwen Edit Plus, Qwen Image 2, Qwen Image 2 Pro, Qwen Image 2.1, Qwen Image 2.1 Pro
Qwen Image is Alibaba’s family of image models. They are known for clear text inside images, in English and Chinese, and for precise photo edits.
Ideogram V4 Turbo, Ideogram V4 Balanced, Ideogram V4 Quality
Ideogram is known for typography: posters, logos and other images where the words have to be right.
Recraft V3, Recraft V4, Recraft V4.1
Recraft makes image models for design work, such as illustrations, icons and brand-style images.
Flux Kontext, Flux Kontext Pro
FLUX Kontext, from Black Forest Labs, edits an image from a plain-language instruction.
FLUX 3
FLUX is Black Forest Labs’ family of image models, made for photorealistic images.
Z Image Turbo, P Image
Pruna AI makes fast image models.
Which model should I use?
- To edit a photo of a person and keep their face: Nano Banana 2 or Nano Banana Pro.
- For a detailed prompt or words inside the image: GPT Image 2.5 or Ideogram V4.
- To combine many reference photos at high resolution: Seedream V4.5.
- For illustrations, icons and design work: Recraft V4.
Every model uses the same credits; see pricing.