Unified image editing and generation






























Qwen Image 2 is a next-generation unified image editing and generation model developed by Black Forest Labs, designed to empower creative professionals with advanced capabilities for transforming, enhancing, and reimagining images. This model stands out by enabling both image generation and detailed image-based editing using intuitive controls, making it especially valuable for artists, designers, filmmakers, and visual content creators who require flexibility and precision.
At its core, Qwen Image 2 allows users to work with one to three reference images, using simple text prompts to instruct how these references should be transformed or combined. Whether you want to update the mood or season of a scene (such as turning a winter landscape into a summery haven), apply stylized transformations, or avoid undesirable elements, the model makes complex image editing accessible through natural language. You can specify desired changes and also set guidance about what you do NOT want to appear in the final artwork.
The system accepts both English and Chinese prompts, providing broad accessibility for global users. Creative control is made approachable through several key user-facing options:
Prompt-based Editing & Generation: Guide the transformation of your reference images using detailed (or brief) descriptions. You simply describe what you envision—for example, asking to "make this park scene appear at sunset" or "add vibrant autumn foliage."
Negative Prompting: Use this feature to steer clear of unwanted content, such as mistakes, visual noise, or off-style artifacts, fine-tuning not only what appears but also what is carefully omitted.
Optimized Prompt Expansion: An AI-powered option helps expand your brief ideas into more detailed prompts behind the scenes, enhancing creative results even if your instructions are short. This is enabled by default for best possible images.
Flexible Image Formats & Sizes: Choose from PNG, JPEG, or WEBP for output, and specify the dimensions to suit your workflow. The model supports a variety of preset sizes (such as square, portrait, and landscape ratios), and can preserve the resolution of your original uploaded images as long as the final image falls between 512x512 and 2048x2048 pixels.
Content Moderation: Built-in safety controls help screen results for inappropriate content, providing peace of mind for professional projects and publisher requirements.
Multiple Outputs in One Go: For fast iteration and creative variety, you can request from 1 to 6 versions of each edit, helping you explore different interpretations or choose the best fit for your project.
Reproducibility: For project consistency or fine-tuned creative workflows, you can generate identical results by fixing the random "seed" value, allowing you to revisit and tweak previous creations reliably.
Qwen Image 2 supports reference images in common professional formats: JPEG, JPG, PNG (without transparency), and WEBP. Input images can be up to 10MB in size and have dimensions ranging from 384 to 5000 pixels on either side, allowing for significant detail. However, output images are constrained to final render sizes between 512x512 and 2048x2048 pixels to ensure optimal quality and performance.
Key creative use cases include visual storytelling, mood or seasonal transformation, scene adaptation for filmmaking and animation, design mockups, stylized reinterpretation, and rapid iteration for content campaigns. The system’s detailed prompt and negative prompt controls allow precise direction, while the ability to process multiple reference images enables composite edits and context-aware transformations.
Best practices include:
In summary, Qwen Image 2 is a strong fit for creative professionals needing a versatile, intuitive, and powerful tool to generate and edit high-quality images based on both visual and textual inspiration.
Add the image that you want change
Přidejte obrázek, který chcete upravit nebo transformovat
A woman kneeling in darkness, illuminated by a warm, radiant beam of light emerging from her raised hand.
Popište požadované úpravy – změny stylu, odstranění objektů nebo vylepšení
Stáhnout profesionálně upravený obrázek
Showcases the model’s ability to dramatically change the season and mood of a landscape photo, opening new creative possibilities for travel, advertising, and film pre-visualization.


Demonstrates advanced time-of-day transformations for architectural photography, allowing users to repurpose images for different moods and marketing needs.


Applies cinematic color grading to landscapes or urban photography, perfect for storytelling, tourism marketing, or creating moody wallpapers with a movie-like atmosphere.


“Transform into a classical oil painting in the style of Rembrandt. Add visible impasto brushstrokes with thick paint texture. Apply warm golden undertones and dramatic chiaroscuro lighting with deep shadows. Enhance the dramatic contrast while preserving facial structure and expression. Add subtle canvas texture visible through the paint layers.”
“Transform into a classical oil painting in the style of Rembrandt. Add visible impasto brushstrokes with thick paint texture. Apply warm golden undertones and dramatic chiaroscuro lighting with deep shadows. Enhance the dramatic contrast while preserving facial structure and expression. Add subtle canvas texture visible through the paint layers.”

Přejděte dnes na syntézu vedenou uvažováním

Image generation with reference consistency
0.2 kredity

Advanced AI-powered image editing
0.4 kredity

Fast intelligent multi-image editor
1.3 kredity

Edit images using text prompts
0.5 kredity

Unified image creation and editing
0.7 kredity

High-fidelity image editing AI
10 kredity

Google's advanced image editing
0.7 kredity

Unified AI image editing model
0.5 kredity

Ultra-fast image editing model
0.1 kredity
Trendující videa