AI startup Midjourney has announced its cutting-edge AI image generation model that enables long tags and significant levels of detail while adding text to images.
The release comes as the company faces increasing pressure from competitors like DALL-E 3, while free Meta tools and services like Leonardo also offer greater control over image creation.
The company says version 6 of its base model delivers improved damage accuracy, improved logical consistency, and finer control of color and shading, as well as making the company's text available for the first time.
The new edition now includes punctuation and grammatical nuances.
“Version 6 allows for precise sequencing prompts, longer prompts, and improved logical consistency and pattern recognition,” Majorney said.
The company highlighted the progress of version 6 compared to the previous version 5.1, which was released in May 2023.
Version 5 - known for its short, easy-to-use tips and great improvements - paves the way for the expanded Version 6.
The ability to create text is one of the standout features of version 6, and while it's not the focus of the template, Midjorny noted that it remains a minor feature.
Although Midjorny's approach to text generation is unique, this feature puts Midjorny in direct competition with other leading models such as Dall-E 3 and Ideogram.
The company describes the feature as a simple function that generates text, saying: "You need to put the text in quotes" because style and aesthetics take precedence over text accuracy.
Version 6 is a little slower and more expensive than Version 5, but the team emphasized that it aims to make the model faster over time.
Version 6 improves zoom in Fine and Creative modes and increases image resolution by 2x.
These features provide users with a variety of creative options, as well as a variety of media to change model resolution, generational differences, and creativity.
These updates show that Midjourney continues to evolve what many consider to be its preferred model and maintains its leadership position even when the company faces challenges from competitors using their own internal models or the popular open source stable deployment model.