Midjourney v6, the latest release in the AI art world, has landed this Thursday. V6 shines with its ultra-realistic imagery and the debut of advanced in-picture text, setting it apart from other diffusion models in generative AI art. The release is shaped by the voices of Midjourney's user community since the users have been involved in the pre-release stage. This underlines Midjourney's unique approach to blending technology with human creativity. In this blog, we'll explore the wows of Midjourney v6, show examples, compare v6 with older versions, and see how to set up and use it.
Better text quality with Midjourney v6
The main drawback of generative AI art has been the mediocre text quality of the images since the beginning of image generation tools. The words on the images usually lacked accuracy and looked visually incoherent. This is due to the complexity of image generation, which requires understanding and replicating language patterns, styles, and legibility in a visual context.
Midjourney v6 impressed many people with the text quality on the images. Many X users have experimented on this and shared the results.
Higher image resolution with Midjourney v6
Midjourney v6 significantly improves the realism of AI-generated images, offering a maximum resolution of 2048 × 2048 pixels, almost double that of its predecessor. With enhanced prompt understanding and the ability to handle over 350 words in prompts, v6 allows users to specify intricate details like colors and object placement. However, it's more sensitive to prompt words, suggesting a shift away from terms like ‘photorealistic, 4k, 8k’ in favor of more straightforward styles.
Photorealism and artistic style of Midjourney v6
Recently, X has been flooded with images generated by Midjourney v6. People are shocked by how realistic, cinematic, and detailed the images are and how well they adapt to the prompt. Here are a few memorable examples.
Midjourney v5.2 vs v6
Here are some side-by-side comparisons of Midjourney v5.2 vs Midjourney v6.
How to use Midjourney v6
Midjourney v6, currently in alpha, is gradually being introduced to its Discord server. Initially, it is accessible to a limited group of users. To use v6, you can change the settings in the Midjourney Discord server or directly message the Midjourney bot. Alternatively, typing “–v 6” after your prompts will also enable the latest version. This method allows users to experiment with the new features of v6.
Midjourney v6 prompting
You'll also need to retrain your prompting skills for Midjourney v6. This model is much better at understanding than the previous ones, meaning you can be more direct and clear with your prompts. For example, if you want the image in a specific style, there's a particular format that you can use instead of prompting with words.
For those aiming for more photographic, less subjective, and literal images in Midjourney v6, the key lies in using the "--style raw" setting. Adjusting the '--stylize' value also plays a crucial role: a lower value (around 100 by default) enhances the AI's understanding of your prompts, whereas higher values (up to 1000) can improve the aesthetics of the generated image. To navigate these new settings effectively, engaging with fellow users on the official Midjourney Discord server can be immensely helpful.
To wrap up, Midjourney v6's launch is a significant milestone in AI art with its striking realism, advanced text integration, image quality, and artistic versatility. These developments not only expand the capabilities of AI tools but also deepen the collaborative relationship between human creativity and machine intelligence. The future of AI art looks promising, with the potential for even more sophisticated and nuanced creations.