Midjourney Unveils New Image Model To Compete With GPT-4o

Initially regarded as one of the leading image generation models in the early AI landscape, MidJourney has seemingly been outpaced by more user-friendly and free tools such as Gemini, ChatGPT, and Bing. The recent upgrade to OpenAI’s GPT-4o model, which offers outstanding image generation and the ability to replicate real photos alongside producing flawless text, has added to MidJourney’s challenges. To remain competitive—especially in light of the Studio Ghibli-inspired AI art trend captivating the internet—MidJourney is introducing a revamped model with numerous enhancements.

CEO David Holz shared insights on the new V7 model via MidJourney’s official Discord server and in a blog post. According to Holz, the new model is “more intelligent with text prompts” and generates images with “significantly increased” quality and “stunning textures.”

The new model is designed to produce images much faster, approximately ten times quicker than the existing version, making it ideal for brainstorming and iterative processes. Users can activate the Conversational mode (accessible only on web) to recreate portions of an image without needing to rewrite the entire prompt or enter Edit mode. These images are of relatively lower quality and only cost half of what regular images do.

One of the most exciting new features for our new V7 model is something we call “Draft Mode.” Draft mode is half the cost and 10 times the speed and it might be the best way to iterate on ideas ever. Try it with voice, think out loud and let our ideas flow like liquid dreams. pic.twitter.com/ANfTMC6Ej1

— Midjourney (@midjourney) April 4, 2025

When using the Discord app on a computer or mobile device, the Conversational mode is replaced by a Voice mode, allowing users to “think out loud” and have images generated fluidly, much like flowing dreams. This function is also integrated into the newly launched Draft mode.

Furthermore, MidJourney V7 can operate in both Relax and Turbo modes, providing high-resolution images (as opposed to Draft mode), with the latter using double the credits for expedited image creation.

At present, the new V7 model is missing some functionalities, and workflows reliant on upscaling, inpainting, and retexturing will revert to the previous V6.1 model. The V7 model also introduces Personalization, allowing users to save their preferences for image styles. This setup process takes around five minutes and involves choosing from a selection of 200 images to refine preferences.

MidJourney is currently conducting a community-driven alpha testing phase for the new model and promises additional features in the upcoming 60 days. To test it out, users can type /settings into the chat box on Discord or the web platform, send the message, and then change the default model to V7 from the available settings.