Midjourney has introduced its latest image-generation model, Version 7 (V7), marking the platform’s first major model release in nearly a year.
Highlights
The new version, which began rolling out in alpha late Thursday (Eastern Time), brings significant architectural changes, enhanced prompt interpretation, and new personalization features aimed at improving the overall image generation experience.
New Architecture and Prompt Understanding
According to Midjourney founder and CEO David Holz, V7 is built on a completely restructured architecture that improves the system’s ability to interpret prompts and produce coherent, detailed visuals.
Enhancements are particularly noticeable in the rendering of textures, hands, bodies, and complex objects—areas that have historically been challenging for AI-generated visuals.
V7 also introduces two operational modes – Turbo and Relax. Turbo mode offers faster generation at a higher cost, while Relax mode trades speed for affordability.
In addition, a new Draft Mode enables image creation at roughly ten times the speed and half the cost of standard generation, albeit at a slightly lower visual quality. Draft images can later be re-rendered or refined for improved output.
Personalization
One of the most notable additions in V7 is its personalization system, which enables users to customize outputs based on their visual preferences.
For the first time, personalization is enabled by default. To activate it, users are prompted to complete a one-time setup involving the ranking of approximately 200 image pairs, allowing the AI to learn their aesthetic style.
Users can further inform the system by liking images on Midjourney’s Explore page. Once this data is collected, personalization can be applied to prompts using the --p parameter.
The intensity of personalization can also be adjusted using the --stylize (or --s) parameter, with values ranging from 0 (no influence) to 1000 (maximum influence), defaulting to 100.
Midjourney also allows users to share and apply personalization codes, enabling others to generate images based on someone else’s stylistic profile by adding the respective code to a prompt.
Performance and Image Quality
Initial results from V7 show improved consistency and alignment between user prompts and generated visuals.
For example, prompts such as “a dragon with a spiky tail” or “a Ghibli-style character” produced outputs that were more visually cohesive and better matched the intended themes.
Although V7 is not specifically optimized for any one artistic style, it appears capable of delivering appealing results across a variety of aesthetics.
Ongoing Development and Limitations
Certain advanced features—including upscaling, retexturing, and zoom-out capabilities—are not yet supported in V7.
Midjourney has indicated that these tools are in development and are expected to be rolled out in the coming months. The current alpha phase invites user feedback to help identify areas for further refinement.
Commercial Growth and Legal Context
Founded in 2022 by David Holz, who previously co-founded Leap Motion, Midjourney remains independently operated and has not received outside funding.
In 2023, reports suggested the company was on track to generate approximately $200 million in annual revenue. Alongside its software initiatives, Midjourney is also building a hardware team and exploring new applications such as 3D and video generation.
However, the company continues to face legal scrutiny over the alleged use of copyrighted material during its training process.
Several ongoing lawsuits accuse Midjourney of using publicly available web images without proper consent. The outcomes of these cases may influence how future versions of the model are trained and deployed.
Practical Benefits of Personalization
The personalization system offers practical applications for a variety of users:
- Brand consistency: Artists and businesses can maintain a cohesive visual identity.
- Creative efficiency: Users can achieve desired results faster without needing to revise prompts extensively.
- Style experimentation: Applying others’ personalization codes allows for stylistic exploration and inspiration.
While Version 7 does not fundamentally change how Midjourney operates, it represents a notable advancement in image quality, user control, and generation speed.
For existing users and newcomers alike, V7 offers a more tailored and efficient creative experience while paving the way for upcoming features.
