Image Generation Models

​

Overview

​

Available Models

​

Flux Models by Black Forest Labs

​

Premium Models

​

Model Comparison

​

Detailed Model Capabilities

​

Flux Schnell

• Quick iterations
• Draft concepts
• High-volume generation
• Time-sensitive projects

• Lightning-fast generation
• Consistent quality
• Low credit cost
• Reliable results

• Less detail than Pro models
• Basic text rendering
• Limited artistic interpretation

"Product photography of a sleek smartphone on white background, 
studio lighting, minimalist style"

​

Flux 1.1 Pro

• Production-ready images
• Marketing materials
• Social media content
• Professional presentations

• Excellent detail
• Strong composition
• Good color accuracy
• Versatile styles

• Moderate generation time
• Less creative than GPT-Image

"Professional business meeting in modern office, diverse team 
collaborating, natural lighting, corporate photography style, 
sharp focus, high resolution"

​

Flux Kontext Pro

• Brand consistency
• Style matching
• Series creation
• Character consistency

Upload a reference image and prompt:
"Generate similar style but with tropical beach setting"

• Maintains visual consistency
• Excellent for variations
• Style transfer capabilities
• Brand alignment

1. Upload brand reference image
2. Generate variations maintaining style
3. Create cohesive visual sets

​

GPT-Image (DALL-E 3)

• Creative concepts
• Artistic interpretations
• Text-heavy designs
• Unique visuals

• Unmatched creativity
• Perfect text rendering
• Complex scene understanding
• Artistic flexibility

• Higher credit cost
• Less photorealistic
• Longer generation time

"A surreal digital painting of a library where books are 
floating and their pages transform into butterflies, with 
text 'IMAGINATION' spelled out in golden letters, dramatic 
lighting, fantasy art style"

​

Google Imagen-4

• Photorealistic images
• Human portraits
• Product photography
• Natural scenes

• Exceptional realism
• Accurate human features
• Natural lighting
• Professional quality

• Longest generation time
• Higher credit cost
• Less artistic interpretation

"Photorealistic portrait of a 35-year-old female CEO in a 
modern office, window light, shallow depth of field, shot 
on Canon 5D Mark IV, 85mm lens, corporate headshot style"

​

Aspect Ratios and Resolutions

​

Supported Aspect Ratios

​

Resolution by Model

• Flux Models: Up to 1024x1024
• GPT-Image: Up to 1792x1024
• Imagen-4: Up to 2048x2048

​

Prompt Engineering Guide

​

Universal Best Practices

1. Structure Your Prompts
   Copy

   [Subject] + [Action/Pose] + [Environment] + 
   [Style/Mood] + [Technical Details]
   
2. Include Technical Terms
   • Camera: “DSLR, mirrorless, iPhone 14 Pro”
   • Lens: “85mm, wide angle, macro”
   • Lighting: “golden hour, studio, natural”
   • Style: “cinematic, minimalist, vintage”
3. Specify Quality Markers
   • “High resolution”
   • “Sharp focus”
   • “Professional quality”
   • “Award-winning”

​

Model-Specific Tips

​

Style Gallery

​

Artistic Styles

​

Advanced Techniques

​

Multi-Image Consistency

1. Use Flux Kontext Pro
   • Upload reference image
   • Maintain style across variations
   • Adjust prompts for different scenes
2. Seed Values (Coming Soon)
   • Lock random seed
   • Generate variations
   • Maintain character features
3. Style Descriptors
   • Define consistent style language
   • Use same technical specifications
   • Maintain color palette

​

Batch Generation Strategy

​

Common Issues and Solutions

​

Poor Quality Results

• Add “high quality, sharp focus, detailed”
• Specify camera/lens for photorealism
• Try a different model
• Avoid conflicting style terms

​

Incorrect Composition

• Specify exact framing: “close-up, full body, wide shot”
• Include compositional rules: “rule of thirds, centered”
• Use photographic terms: “portrait orientation, landscape”

​

Style Mismatch

• Use Flux Kontext with reference
• Be more specific about style
• Include artist or movement references
• Remove conflicting descriptors

​

Text Rendering Issues

• Use GPT-Image for text
• Put text in quotes: “SALE”
• Specify font style
• Keep text simple

​

Credit Optimization

​

Cost-Effective Strategies

1. Tiered Approach
   • Concepts: Flux Schnell (5 credits)
   • Drafts: Flux Pro (5 credits)
   • Finals: Premium models (10 credits)
2. Batch Similar Requests
   • Group by style
   • Use consistent prompts
   • Leverage successful formulas
3. Smart Model Selection
   • Flux for general imagery
   • GPT-Image for creative/text
   • Imagen-4 only for photorealism

​

ROI Calculation

Example Project: 10 marketing images
- Testing: 20 x Flux Schnell = 100 credits
- Refinement: 10 x Flux Pro = 50 credits
- Finals: 10 x GPT-Image = 100 credits
Total: 250 credits (vs 500 if all premium)

​

Integration Tips

​

With Workflows

• Idea to Video: Flux Pro for efficiency
• Script to Video: Match model to content type
• Marketing: GPT-Image or Imagen-4
• Educational: Flux Pro for clarity

​

With Brand Kit

• Use Kontext Pro with brand references
• Maintain consistent style
• Apply brand colors in prompts
• Include logo placement instructions

​

Next Steps