# Choosing the right model Chatly offers AI models for **image generation** and **editing**, each built for different creative tasks. Some focus on creating new images, others on refining existing ones, and some can do both. Knowing what each model does best helps you get the results you want faster. ### Auto Mode **Auto** automatically selects the best model based on your prompt. It analyzes the task you want to perform, whether creating a new image, editing an existing one, or both, so you do not have to choose a model manually. **New Image Generation:** If your prompt is purely for new image creation, Auto will send it to the most relevant generation-capable model according to your task. This ensures every request is handled by the most suitable model without you needing to switch manually. **Editing Requests:** If you’ve chosen a model that only supports image generation but your prompt includes editing instructions (for example, *“change the background”* or *“add a new object”*), Auto will **automatically reroute** your request to an editing-capable model so your task completes successfully. This intelligent routing allows you to focus entirely on your creative process while the system handles model selection. ### Available Models Chatly offers multiple AI models designed for both **image generation** and **image editing**. Each model has its own strengths and ideal use cases. Some specialize in creating completely new visuals from text, others focus on detailed image editing, while a few can handle both. ### Model Categories and Use Cases #### 1. Dual-Use Models (Generation + Editing) #### Nano Banana Models **Nano Banana** **Capabilities:** Supports both creating new images and editing existing ones. **Best For:** * Merging input images at a time into a single visual (lifestyle composites, ad visuals) * Adjusting lighting, colour, or exposure * Cleaning up, retouching, or removing unwanted objects * Creating visuals and refining them iteratively **Applications:** * Product photography and branding * Fashion or character design * Home or interior visualisation * Marketing visuals requiring iteration and refinement **Who Should Use It:** Designers, marketers, content creators, and anyone who needs a single model for merging and editing images {% embed url="" %} **Nano Banana Pro** **Capabilities:** A more powerful version of Nano Banana with higher resolution, advanced text rendering, and high-accuracy merging. **Best For:** * Merging multiple photos and elements seamlessly * Diagrams and technical visuals with precise text * Posters, graphics, and layout-heavy designs * Professional photo editing and upscaling * Maintaining character consistency across multiple images **Applications:** * High-quality brand visuals and ad graphics * Complex diagrams, labelled images, educational content * Product and lifestyle images in high resolution * Upscaling or restoring old/low quality photographs **Who Should Use It:** Designers, educators, brands, marketers, and creators who need top-tier quality with reliable text and merging. {% embed url="" %} #### Seedream Models **Seedream 4.0** **Capabilities:** Generating and editing images in 4K quality across multiple use cases and categories. **Best for:** Characters, product mockups, illustrations, wallpapers, portraits * Produces **stylized, creative, and textured images** * Ideal for concept art, editorial visuals, or fantasy work * Great for expressive characters and detailed scenes **Applications:** * Character creation for games, comics, or stories * Product mockups or stylized marketing visuals * Wallpapers, digital paintings, and illustrations * Portraits with creative lighting and textures **Who Should Use It:** Digital artists, photographers, and creatives who want expressive, original visuals **Note:** *Images generated with Seedream cannot be upscaled since they are already in 4k quality.* **Seedream v4.5** **Capabilities:** Generating and editing images in ultra-clear 4K quality with strong detail, consistency, and creative control. **Best for:** Characters, product mockups, illustrations, wallpapers, portraits * Produces high-fidelity, stylized, and detailed images * Handles consistent characters, products, and scenes across multiple images * Strong with visuals that include text, labels, posters, and layout design **Applications:** * Visual series that require the same character or product across multiple angles * Designing characters for storytelling, animation, or worldbuilding * Producing polished product visuals or campaign-ready mockups * Creating digital art, stylized posters, or full-scene illustrations * Generating portraits with dramatic lighting, mood, and fine detail * Building cohesive multi-image sets for branding or creative projects **Who Should Use It:** Artists, designers, brands, and creators who want high-quality, highly detailed images with strong style control and consistency. **Note:** *Images generated with Seedream 4.5 do not require upscaling, as they are already rendered in high-resolution 4K.* {% embed url="" %} **Flux 2 Pro** **Capabilities:** Powerful dual-purpose model for generating new images **and** editing existing ones with high precision. **Best For:** * Products, homes, portraits, wallpapers * Food photography and restaurant visuals * Movie-style posters with dramatic lighting **Applications:** * Product photos that need clean adjustments * Interior or architectural improvements * Portrait retouching or creative styling * Food and menu visuals * Cinematic poster concepts **Who Should Use It:** Designers, marketers, photographers, and creators who want a strong all-rounder for editing + generation in one place. {% embed url="" %} **Z Image Turbo** **Capabilities:** Generates photorealistic images quickly, with sharp detail, realistic textures, and accurate text rendering. **Best for:** Product visuals, marketing assets, posters, social media graphics, concept art * Produces high-quality images in seconds * Maintains realistic lighting, texture, and fine detail * Handles text accurately for labels, captions, and graphics **Applications:** * Product photography and e-commerce visuals * Marketing campaigns, banners, and social media posts * Posters and promotional graphics * Concept art or rapid prototyping of visual ideas * Projects requiring fast iteration with precise text placement **Who Should Use It:** Designers, marketers, content creators, and anyone who needs fast, realistic images with accurate text and clean details. {% embed url="" %} #### 2. Image Generation Models **Ideogram 3.0** **Capabilities:** Strong text rendering and layout design **Best for:** Posters, social media visuals, cards, ads, logos, and typography * Great for structured layouts and text integration * Keeps designs clean and professional * Works well for marketing campaigns, branding, and editorial visuals **Applications:** * Ads, banners, and social media posts * Logo or branding concepts * Posters and card designs **Who Should Use It:** Designers, marketers, and anyone who needs polished, production-ready graphics {% embed url="" %} **Qwen Image** * **Capabilities:** Generates detailed, natural-looking images with controlled composition **Best For:** * Character concepts and design * Wallpapers and abstract designs **Applications:** * Storyboards or design reference boards * Creative concept art * Character design for movies and games **Who Should Use It:** Storyboard artists, visual and game designers {% embed url="" %} #### 3. Image Editing Models **Flux Kontext Pro** * **Capabilities:** Specialised in advanced, fine-grained image editing **Optimised for:** Advanced edits, object changes, colour adjustments, environment modifications * Strong context awareness for realistic edits * Handles complex changes without breaking the image **Applications:** * Move, replace, or remove objects * Adjust colours, lighting, and shadows * Change backgrounds or interiors * Professional retouching * Creative transformations of existing images **Best for:** Designers, photographers, and teams needing high-quality, precise edits ***Note**: Flux Kontext Pro can only take 1 input image. If 2 images are uploaded, the model will automatically route to Nano Banana.* {% embed url="" %} **Qwen Image Edit** * **Capabilities:** Focused on localised image adjustments and quality refinements **Optimised for**: Subtle edits, refinements, localised adjustments, retouching * Perfect for fine-tuning details * Preserves the original style while improving quality **Applications:** * Smooth textures or fix small flaws * Add or remove small elements * Maintain style consistency * Minor composition tweaks **Best for:** Creators, marketers, or designers who want precise, controlled edits {% embed url="" %}