Midjourney, one of the first platforms to use generative AI for image creation, has introduced a new model called V7. This is the company’s first new release in almost a year. The new version is already available in alpha and offers users improved image quality and the ability to customize results based on personal preferences.
To use the new model, users need to go through a personalization process by rating about two hundred images. This creates a profile that tunes V7 to individual visual preferences. The model can be activated on the Midjourney website or via the Discord chatbot. Users can select the model from a dropdown menu next to the “version” label.
According to Midjourney CEO David Holz, V7 has “an entirely different architecture” and demonstrates significantly improved ability to process text prompts. The image quality is noticeably higher, with enhanced textures and object detail. The model is available in two modes — Turbo and Relax — and supports a new tool called Draft Mode, which allows images to be generated ten times faster and at half the cost of the standard mode.
For now, some standard Midjourney features, such as upscaling and retexturing images, are not available for V7. However, the company promises to add them within the next two months. Developers encourage users to test the new model and share their feedback to help identify its strengths and weaknesses.
