KREA

Midjourney is an AI image generation service that creates artwork from text prompts. Users describe what they want to see, and Midjourney produces four image variations to choose from. The platform is known for its distinctive artistic style, producing images with rich textures, dramatic lighting, and a painterly quality that sets it apart from competitors. Beyond basic generation, Midjourney offers style reference, character reference, regional variation, and remix modes for refining results. Originally accessible only through Discord, it now has a full web interface and has expanded into video generation.

DALL-E 3 is OpenAI's text-to-image model that generates detailed artwork from natural language prompts. Released in October 2023, it represents a significant improvement over earlier versions, following complex prompts with better accuracy and producing more coherent images. It can generate images in various styles including photorealistic, illustrative, and emoji-like. DALL-E 3 is integrated into ChatGPT Plus and Enterprise for direct generation, and available through the OpenAI API for developers building image generation into applications. It also supports inpainting, outpainting, and variations for modifying existing images.

Stable Diffusion is an open source text-to-image model that generates detailed pictures from text descriptions. Released in 2022 by Stability AI, it democratized AI image generation by making the model weights and code publicly available, allowing anyone to run it on their own hardware. Beyond basic text-to-image generation, it supports inpainting, outpainting, image-to-image translation, and depth-aware generation. The ecosystem includes ControlNet for precise control, DreamBooth for personalization, and interfaces like ComfyUI and AUTOMATIC1111. The latest version, SD 3.5, uses a multimodal diffusion transformer architecture for improved quality.