Midjourney, a leader in artificial intelligence image generation technology, has released a significant update with version 6.1, enhancing the realism of human skin textures and the clarity of text in images. This update moves away from the anticipated release of version 6.5 and instead continues an iterative update path with version 6.2 expected soon.

What’s new?

The most substantial improvements in version 6.1 concern the depiction of human figures. Adjustments have been made to the portrayal of arms, legs, hands, and overall body shapes. The update introduces advanced texture mapping techniques that offer more sophisticated skin textures, significantly enhancing the lifelike quality of images. 

Midjourney version 6.1 brings comprehensive enhancements across the model, improving the overall image quality by reducing pixel artifacts, refining texture details, and optimising the rendering of various artistic styles, including 8-bit and retro. The company’s announcement on X reflects these improvements: “Version 6.1 significantly boosts image quality, coherence, and text clarity, alongside introducing new upscaling and personalization models. It is smarter, faster, clearer, and more aesthetically pleasing. This is, without doubt, our finest model to date.”

Upscaling and Textural Improvements

A notable feature of the update is the improved upscaling process, which now delivers higher-quality images and textures, enhancing the overall visual experience. The new “–q 2” mode, though requiring more processing time, adds considerable texture to the images, improving their realism.

The update also enhances precision in depicting small image features, which is especially beneficial for accurately representing details like eyes, small faces, and distant hands. Furthermore, Midjourney has tackled one of the common challenges in AI image generation – text accuracy. 

How to Use Midjourney Version 6.1

To access the enhancements of Midjourney version 6.1, users need to add “–v 6.1” to their image generation prompts. This command is compatible with both the web and Discord versions of the platform.

Another update is that your –p code is changed with 6.1. It still makes beautiful results though!

Testing and prompting

If you’ve followed our content a little, you’d have very quickly realised I’m into AI legislation and art. And that I like my art darker than your soy sauce or black coffee. So, in a darker lens, let’s compare some images.

Metropolis inspired dystopian images, compared with Midjourney v.6 (left) and v6.1(right), no personalisation included:

Midjourney v.6 - Metropolis inspired black and white image
Metropolis inspired black and white image

While testing some more glitchy methods, I got very different results using the same prompt. The left image is Niji 6 and kind of gives me nightmares, the other one has this ethereal realism of disintegration in a digital realm. Definitely not what I expected, but I’m satisfied with the results. 

Niji 6 - Glitch art
Midjourney v 6.1 - Glitch art
Testing with personalisation

Let’s compare Midjourney 6.0 without –p (left), 6.0 with –p (middle) and 6.1 with  –p included (right), with the same prompt. I can definitely say I notice less weird glitches and the image is overall more coherent. The ray tracing is also improved, but I haven’t tested this enough. 

Midjourney 6.0 without --p Watching soul sucking TV
Watching soul sucking TV - Midjourney 6.0 with --p
Watching soul sucking TV - Midjourney 6.1 with --p

Playing with some iridescence for the next examples, left Midjourney 6.0 and right 6.1, both with –p  included. The skin texture on the 6.1 model is incredibly detailed, right to the skin cells!

Crystalline portrait - Midjourney 6.0
Midjourney 6.1 - Crystalline portrait

Now my favoute image, however probably the darkest of them all. Midjourney v6 is on the left and 6.1 on the right. This image resonates with me for personal reasons, but art is subjective, isn’t it?

Ribcage of flowers
Ribcage of flowers
Thoughts and opinions

Overall, Midjourney’s update didn’t fail us. I’m excited about what other updates will bring. I’m amazed by the texture, even with blended personalised images that usually come out in lower quality. Haven’t tried the text feature yet, since I use other softwares for design, but I’ll post an update. 

How has your experience been with the new features?  

Since my art style is unusual, is there a prompt, style or idea you want me to try? Feel free to share your thoughts and feedback – I’d love to hear how others are using these enhancements in their creative projects.