Midjourney Releases V7, Its First New AI Image Model In Nearly a Year

Midjourney has released its new V7 image model, which features a revamped architecture with improved text prompt handling and higher image quality. The model also includes default personalization based on user-rated images, which requires rating around 200 images to build a personalization profile. This profile tunes the model to individual visual preferences, making it the first Midjourney model to have personalization switched on by default. The V7 model can be accessed on Midjourney's website and Discord chatbot, and can be selected from the drop-down menu next to the "Version" label. According to Midjourney CEO David Holz, V7 is a "totally different architecture" that is "much smarter with text prompts" and produces higher-quality images with better coherence on details. The model comes in two flavors, Turbo and Relax, and powers a new Draft Mode that renders images at 10x the speed and half the cost of the standard mode. Draft images are of lower quality but can be enhanced and re-rendered with a click. Some standard Midjourney features, such as image upscaling and retexturing, are not yet available for V7 but are expected to arrive in the near future. Holz encourages users to play around with the new model to learn its strengths and weaknesses. Overall, V7 promises to offer improved image quality and personalization, but may require different styles of prompting.