Midjourney Unveils Version 6
The image generation AI model Midjourney released a latest update V6, introducing in-image text and refined prompting features.
Midjourney has been a pioneering artificial intelligence (AI) tool for image creation, and its latest offering, version 6 (V6) , marks another significant step in the development of AI-driven image generation. Released as an alpha version, Midjourney V6 arrives with a suite of enhanced features, most notably the ability to generate in-image text and an overhaul in its prompting mechanism.
Founded by David Holz, the Midjourney research collective has been at the forefront of AI image generation. The new features in V6 include more realistic and detailed image generation, a capability that has set Midjourney apart in the competitive landscape of AI art tools. This advancement addresses one of the most sought-after aspects in AI imagery, especially in comparison to its counterparts like OpenAI’s DALL-E 3 and Ideogram.
Holz, in a message to the Midjourney Discord community, which boasts over 17 million members, highlighted the significant improvements in V6. He described it as the third model developed from scratch, leveraging their AI superclusters. V6’s capabilities extend to more accurate prompt following, allowing for longer prompts and improved coherence. The model knowledge has seen enhancements, along with the image prompting and remix features.
A notable addition is the minor text drawing ability, where users can input text in quotations to generate images with embedded text. This feature is particularly innovative, adding a new dimension to AI-generated art. The new version also introduces improved upscalers with subtle and creative modes, enhancing image resolution.
Additionally, Holz mentioned that V6 requires a new approach to prompting. The previously used technical and esoteric methods, like mentioning camera names or film stocks, will no longer yield the desired results. This shift necessitates users to re-learn how to prompt, focusing on being explicit and avoiding unnecessary additions to their prompts.
Enthusiastic users, including horror director and digital artist Chris Perna, have started exploring V6, sharing vibrant and richly detailed results on social media platforms. The text generation feature, in particular, seems to hold a lot of promise.
Despite these advancements, V6 is still an alpha test, as noted by Holz. Users can expect frequent changes and updates as it progresses towards a full release. Many visual content creators have been employing a variety of AI tools in their work, and creating images with Midjourney is often an initial step into an artwork or video. And thus these updates can have further influence in the broader creative and art industries.
Midjourney's V6 underscores the company's position in the AI art generation field, innovating and advancing despite the competitive environment and the challenges posed by other models like Stable Diffusion. However, Midjourney and similar AI art generators are currently embroiled in copyright infringement litigation, with artists accusing them of using their publicly posted work without consent. The outcome of this legal battle, particularly regarding the fair use defense, remains to be seen.