Midjourney Unveils First New Model In A Year
AI image generator Midjourney has unveiled Version 7 (V7), its first new model in over a year. Now available in public alpha, V7 brings a host of improvements, with personalization enabled by default. The company describes this update as a complete rework rather than a simple upgrade.
The new model enhances image generation in two core areas: visual quality and creative workflow. With V7, users will immediately notice greater accuracy when it comes to details like hands, body parts, and intricate textures. While past versions struggled with producing realistic human features, V7 handles them with much more coherence. Moreover, materials such as skin, metal, and ceramics appear more lifelike, thanks to refined texturing and lighting effects.
Although AI-generated images are still detectable by experts, this version reduces many of the visual quirks that made earlier outputs stand out. V7 can now reproduce complex lighting scenarios and material nuances that its predecessor, V6.1, often failed to achieve.
New Features and Performance Modes
One of V7’s most notable additions is “Draft Mode.” Midjourney states that this mode renders images at ten times the speed and at half the cost compared to standard generation. Despite this, draft images are lower in quality and intended for testing concepts rather than final output. This allows users to iterate rapidly before settling on a final design.
In addition to Draft Mode, V7 includes two generation speeds: Turbo and Relax. Turbo creates high-quality images faster but consumes more credits. Conversely, Relax Mode processes images slowly but is more cost-effective. At launch, V7 lacks a standard mode, though Midjourney assures users it will arrive once further testing is complete.
The model also supports familiar parameters like --ar and --seed, along with the personalization system first introduced in V6.1. Unlike before, personalization is now active by default. To use it, users must select at least 200 images to train the model in identifying their preferred aesthetic. However, this feature remains optional and can be disabled anytime.
Evolving Platform and Legal Concerns
Initially accessed through Discord, Midjourney has since expanded to a user-friendly web interface. It continues to be a staple tool among digital artists and video creators. Often, creators start their projects in Midjourney before refining them using video tools like Runway.
Despite its popularity, Midjourney remains entangled in legal scrutiny. Critics argue that training AI models on copyrighted material may violate intellectual property laws. This concern is not without merit—some Midjourney images still include remnants of watermarks and artist signatures.
Recently, the company hinted at future plans to develop hardware, though no specific details have been shared. While V7 marks a significant step forward for AI-generated imagery, the legal and ethical implications of its training methods are likely to remain in focus.
