Midjourney, a pioneer in AI image generation, has released its first new model in nearly a year, dubbed V7. The model started rolling out as an alpha version last week, following OpenAI’s debut of a similar image generator in ChatGPT.
V7 is the most advanced and beautiful model yet, according to Midjourney CEO David Holz. It boasts improved text prompts, higher image quality with detailed textures, and better coherence on all details. The model comes with two flavors, Turbo and Relax, which offer different performance and costs.
One notable feature of V7 is its built-in personalization profile, which tunes the model to individual visual preferences through a rating system that involves around 200 images. This allows users to quickly select the model from the drop-down menu in the web app or Discord chatbot.
Holz described V7 as a “totally different architecture” and emphasized its unique strengths and potential weaknesses. He encourages users to provide feedback on the model’s performance, suggesting that it may require different styles of prompting.
In recent months, Midjourney has been working on expanding its capabilities, including image upscaling and retexturing, which are expected to be available soon. The company remains committed to developing innovative AI tools without relying on outside funding, having raised no external money since its founding in 2022.
The release of V7 comes amidst some challenges for Midjourney, as the company faces several lawsuits alleging infringement on artists’ rights due to its use of scraped images without consent. Despite these issues, the development of V7 marks an exciting milestone for the AI image generation space.
Source: https://techcrunch.com/2025/04/03/midjourney-releases-its-first-new-ai-image-model-in-nearly-a-year