After almost a year in development, Midjourney has finally released its latest version: Midjourney v7. This alpha version brings numerous improvements and innovative features that take AI image generation to a new level. In this blog post, we take an in-depth look at the most important innovations and what they mean for creatives.
The most important innovations at a glance
Improved image quality and texture representation
Midjourney v7 offers significantly higher image quality with more realistic textures and better coherence in the representation of bodies, hands and objects. The AI now understands text prompts much better, which leads to more precise results, and the interpretation of image prompts has also been significantly improved. Even subtle details such as skin wrinkles or the surface structure of materials are now more authentically represented – aspects that were often considered weak points in previous AI image generators.
Personalization as standard
A significant new feature is the model personalization that is activated by default. To use v7, users must first evaluate about 200 image pairs, which takes about five minutes. This process enables the AI to learn the user’s aesthetic preferences and interpret prompts more individually. After this one-time setup, personalization can be turned on or off at any time.
Draft Mode: Iterate faster and cheaper
The new “Draft Mode” is one of the most exciting features of v7. It generates images ten times faster than the standard mode and costs only half as much. Although the images are created in lower quality, behavior and aesthetics remain consistent, making the mode ideal for quick brainstorming and idea generation.
In the web interface, the input line in Draft Mode transforms into a “conversational mode” that allows users to make adjustments to images without having to reformulate the entire prompt. Particularly interesting: Draft Mode also supports voice input via a “Voice Mode” function that can be activated by clicking on the microphone icon.
Two speed modes: Turbo and Relax
Midjourney v7 is offered in two speed variants:
- Turbo: Faster, but twice as expensive as a standard v6 job
- Relax: Slower, but more cost-effective
A standard speed mode for v7 is in development and is expected in future updates.
What’s missing and what’s coming
Midjourney plans to release new features every one to two weeks over the next 60 days. The biggest upcoming feature will be a new v7-specific reference system for characters and objects – a potentially powerful tool for consistent representations across multiple images.
CyberJungle did a great summary
Mit dem Laden des Videos akzeptieren Sie die Datenschutzerklärung von YouTube.
Mehr erfahren
Bottom line: a significant leap forward
Midjourney v7 represents a significant advance in AI image generation. With improved image quality, a personalized experience, and innovative features like draft mode and voice control, it offers creatives new ways to realize their ideas faster and more efficiently.
The release comes just one week after OpenAI presented its new image generator in ChatGPT, which attracted a lot of attention for its ability to generate Ghibli-style images. In this competitive environment, Midjourney is once again at the forefront of AI image generation with v7.
Sources:
https://www.midjourney.com/updates
