Supported Image Models

Chat Video Pro supports multiple state-of-the-art AI image generation models, each optimized for different use cases, quality levels, and workflows.

Model Comparison

Model	Resolution	Best For	Special Features
Flux 2 Max	Up to 1536×1536	Highest quality, detailed images	Text-to-image, edit variant
GPT Image 1.5	1024×1024	Canvas editor, composition	Advanced editing, layers
Nano Banana Pro	Up to 1536×1536	Fast, high quality	14 image inputs, versatile
Z-Image Turbo	Up to 1024×1024	Speed, quick iterations	Fast generation
Seedream v4.5	Up to 1024×1024	Creative, artistic	Text-to-image, edit variant

Quick Selection Guide

Choose Flux 2 Max if:

• ✅ You need the highest quality images

• ✅ You want maximum detail and realism

• ✅ You're creating final production assets

• ✅ Budget allows for premium model

Choose GPT Image 1.5 if:

• ✅ You need Canvas Editor features

• ✅ You're doing complex composition work

• ✅ You need layer and mask support

• ✅ You want advanced editing capabilities

Choose Nano Banana Pro if:

• ✅ You need to work with multiple images (up to 14)

• ✅ You want fast, high-quality results

• ✅ You need versatility across use cases

• ✅ You want good balance of quality and speed

Choose Z-Image Turbo if:

• ✅ Speed is your priority

• ✅ You're iterating quickly

• ✅ You need fast previews

• ✅ Quality is less critical than speed

Choose Seedream v4.5 if:

• ✅ You want creative, artistic results

• ✅ You're exploring different styles

• ✅ You need both text-to-image and edit capabilities

Model Variants

Many models have two variants:

Text-to-Image Variants

• Purpose: Create images from text prompts

• Use when: Starting from scratch

• Examples: flux-2-text-to-image, seedream-v4-text-to-image

Image-to-Image (Edit) Variants

• Purpose: Edit and modify existing images

• Use when: You have an image to modify

• Examples: flux-2-edit, seedream-v4-edit

• Auto-switching: System automatically switches when you attach images

Canvas Editor Support

Some models support the advanced Canvas Editor:

Model	Canvas Support	Best For
GPT Image 1.5	✅ Full support	Composition, layers, masks
Nano Banana Pro	✅ Full support	Multi-image composition
Flux 2 Max	✅ Full support	High-quality editing
Seedream v4.5	✅ Full support	Creative editing
Z-Image Turbo	✅ Full support	Quick edits

Input Limits

Different models support different numbers of input images:

Model	Max Images	Use Case
Nano Banana Pro	14 images	Complex compositions
GPT Image 1.5	4 images	Canvas editor workflows
Flux 2 Max	1-2 images	Standard editing
Seedream v4.5	1-2 images	Standard editing
Z-Image Turbo	1 image	Quick edits

Resolution Support

Model	Max Resolution	Aspect Ratios
Flux 2 Max	1536×1536	16:9, 9:16, 1:1, 4:5, 21:9
GPT Image 1.5	1024×1024	1:1 (square)
Nano Banana Pro	1536×1536	16:9, 9:16, 1:1, 4:5
Z-Image Turbo	1024×1024	16:9, 9:16, 1:1
Seedream v4.5	1024×1024	16:9, 9:16, 1:1

Quality vs. Speed

Premium Quality (Slower)

• Flux 2 Max - Highest quality, slower generation

• GPT Image 1.5 - High quality, advanced features

Balanced

• Nano Banana Pro - Good quality, reasonable speed

• Seedream v4.5 - Creative quality, moderate speed

Speed Optimized

• Z-Image Turbo - Fast generation, good quality

Special Features

Canvas Editor Integration

• GPT Image 1.5 - Best for complex composition

• Nano Banana Pro - Best for multi-image workflows

• Flux 2 Max - High-quality editing

Multi-Image Support

• Nano Banana Pro - Up to 14 images

• GPT Image 1.5 - Up to 4 images

• Others - 1-2 images

Background Removal

• Specialized models available

• See Background Removal for details

Upscaling

• Separate upscaling models

• See Image Upscaling for details

Model Selection Tips

1. Start with Nano Banana Pro - Good balance for most use cases

2. Use Flux 2 Max for final assets - When quality is critical

3. Choose GPT Image 1.5 for Canvas - When you need advanced editing

4. Use Z-Image Turbo for speed - When iterating quickly

5. Check input limits - Ensure model supports your image count needs

Auto-Switching Behavior

The system automatically switches between text-to-image and edit variants:

• No images attached → Text-to-image variant

• Images attached → Edit variant (if available)

• Manual override → You can manually select variants

Next Steps

• Learn about Text-to-Image for creating from scratch

• Explore Image-to-Image for editing existing images

• Check out Canvas Editor for advanced editing