Midjourney has begun alpha testing its highly anticipated V7 model, introducing significant improvements such as realistic human hand images, along with other enhanced features aimed at delivering more accurate and detailed AI-generated artwork
Enhanced image quality and personalisation
The V7 model boasts improved image quality, particularly in rendering human bodies, hands, and objects with greater coherence and realism.
Midjourney’s team notes that the model is “much smarter” at processing text prompts, allowing for more accurate and detailed results based on user input.
By default, personalisation allows users to fine-tune outputs based on their visual preferences. Users can create their personalised style in about five minutes and toggle it on or off as needed.
Midjourney’s V7 addresses long-standing issues in AI-generated imagery, particularly the accurate depiction of hands, which have historically been challenging for AI models.
According to some users, V7 can produce photorealistic images of human hands, showcasing its enhanced capabilities.
Draft mode and operational enhancements
One of the standout features of the V7 model is its “Draft Mode,” which operates at half the cost and produces images ten times faster than standard modes.
This mode is perfect for exploring variations or quickly iterating on ideas. Additionally, users can utilise voice commands in Draft Mode, making it easier to express their thoughts without typing.
Additionally, V7 includes Turbo and Relax modes for standard rendering, with Turbo costing twice as much as a regular speed job on the V6 model.
While some features, such as upscaling, editing, and retexturing, will initially rely on the V6 model, Midjourney plans to roll out updates every one to two weeks over the next sixty days.
As users start testing the V7 model, Midjourney welcomes feedback to refine its capabilities further. The community’s contributions will be vital in shaping the model’s final version.