Midjourney is in the final testing phase of V7, marking a significant update that has overhauled nearly every aspect of the AI image generation system.
During this week's office hours meeting, developers explained that V7 features a new architecture, different datasets, and revised language processing methods.
Internal testing could take anywhere from one week to a month, followed by a two-week debugging phase and potentially a limited-functionality alpha stage.
"The V7 model represents a complete overhaul of V6, with new architecture, different datasets, and revised language processing methods," announced Midjourney on its official server. "There is usually an alpha phase after initial debugging, with limited enabled functionalities."
While extensive refinements for V7 continue, Midjourney's team is also exploring video generation capabilities through various approaches, including developing their own video models or collaborating with external providers.
The main challenge lies in making video creation both user-friendly and cost-effective, according to the team.
The team has not disclosed a specific timeline for the video generator, suggesting it may still be in early development stages. Currently, developers are focusing on balancing speed, cost, and output quality based on announcements.
Around 85% of users prefer images generated by the platform's personalized system, which now includes mood boards and multiple profiles, Midjourney reports. The company plans to enhance these features by integrating mood boards with style reference tools.
The platform also intends to introduce two distinct generation modes: a "live" option for quick results similar to Meta's "Imagine" feature, Krea AI's doodle-to-image function, or Leonardo's live canvas; and a high-volume mode for creating large batches of images akin to Leonardo's Flow state.
Midjourney V6 was released in December 2023, offering better results than its predecessor, V5.1, launched in May 2023.
The improvements were notable, though the upgrade from V5 to V6 wasn't as substantial as the leap from V4 to V5. Given Midjourney's cautious tone when discussing V6, it's reasonable to infer that the team is highly satisfied with the new version, describing it as a "complete overhaul."
During this period, the team also announced they were entering the AI hardware market but provided no further details.
Since the release of Midjourney V6, the team has expanded its functionality by adding image editors, standalone interfaces, style customization tools, outpainting, and image expansion features.
The team has also started gathering user feedback through various surveys to understand what features users would like to see implemented.
According to Decrypt's analysis, Midjourney excels in creative AI image generation, consistently producing aesthetically pleasing results that set it apart from competitors.
Though it may occasionally sacrifice strict adherence to prompts and text rendering accuracy, its versatility and overall capabilities make it a powerful tool.
The addition of customizable styles further enhances its appeal, allowing users to create unique personalized images that differ significantly from the platform's signature look.
