Contents
AI image generator MidJourney has released version 6 of its popular tool, allowing you to make longer tooltips, increase the level of detail, and even add text to your photos.
MidJourney first entered open beta in July 2022 and has made rapid progress in image quality and accuracy with each subsequent release. This release comes as MidJourney faces increasing pressure from competitors, particularly OpenAI's DALL-E 3. Likewise, free tools from Meta and services like Leonardo offer an increasing number of image generation options.
What's new in MidJourney v6?
The latest release of MidJourney increases tooltip length, provides more granular control over color and shading, text, and provides the ability to fine-tune the output during a conversation with MidJourney, much like ChatGPT allows you to enhance the DALL-E image.
The latest update has been in development for some time, with separate groups of users testing and comparing results to improve the overall quality. One of the most striking updates concerns the way clues are interpreted and understood. For example, the tool can now understand the nuances of punctuation and grammar, reports Toms Guide.
Poster with text
MidJourney v6 was asked to create a poster for the movie “Puppies of the Moon”, but the program was unable to add text to the image. When the program was asked to add certain words by putting them in quotes, it solved the problem.
Comparison with previous version
MidJourney v5.2 (left) and MidJourney v6 (right) generated an image from the same clue “Wine taster looking through glass.”
Vacancies
Journalist, author of stories about IT, business and people in MC.today MC.today
Junior Data Analyst TRAFFBRAZA
Senior Account manager IT Delight, Kiev
Content Marketing Manager YozmaTech, Viddaleno, salary 1500
Real people
Some AI models refuse to create images of a living person, including DALL-E, but MidJourney does not experience these problems.
v5.2 (left), and v6 (right) created an image of Elon Musk riding a unicorn. None of them received quick revision or follow-up instructions. Version 5.2 created a cartoon-like image (and no angle), while version 6 looks like a stylized photograph.
