Midjourney v5 is now available – check out the changes from v4!
Midjourney v5 has finally arrived and we're excited to share with you the latest updates and enhancements from previous version, v4.
Whether you're a frequent user or new to Midjourney, this upgrade promises to take your journey planning experience to the next level.
So fasten your seatbelt and join us as we explore what's new in Midjourney v5!
Introducing Midjourney v5
As part of its commercial AI image-synthesis service, Midjourney announced version 5, which can produce photorealistic images of the highest quality, which some AI art fans are calling creepy and "too perfect."
With Midjourney v5, the popular AI image generator's neural network can generate strikingly lifelike visuals. The v5 model uses cutting-edge tools and a new neural architecture to generate aesthetics and designs.
Customers who subscribe to Midjourney's Discord service can now test Midjourney v5 as an alpha release.
"According to the announcement, these are the changes that come with the update:
- Much wider stylistic range and more responsive prompting.
- Much higher image quality (2x resolution increase) improved dynamic range.
- More detailed images. Details are more likely to be correct. Less unwanted text.
- Improved performance with image prompting.
- Supports --tile argument for seamless tiling (experimental).
- Supports --ar aspect ratios greater than 2:1 (experimental).
- Supports --iw for weighing image prompts versus text prompts."
More realistic output images
In addition to generating more realistic and detailed images, the new model also requires more precise instructions.
A new neural architecture and new aesthetic techniques have been used to develop the model, which was trained on an "AI supercluster" in the Google Cloud for about five months, according to Midjourney. This model is able to produce more realistic images and more accurate details.
In a first test, the same prompts without explicit image styles resulted in more detailed images with significantly more realistic results. In this example, Midjourney v4 produces an illustration, whereas v5 produces a more photorealistic installation in a museum for the prompt “a tree made of money”. As a result of v5, the result is also much closer to its intended outcome.
Sharper, more detailed images
Since upgrading Midjourney, you'll notice that your images are now sharper and more detailed. This is thanks to the new image processing algorithms that have been implemented in this latest release.
As you can see in the comparison between V4 and V5, V5 produces sharper, more detailed images. Here’s a comparison of V4 vs. V5 for the prompt: “photo of two toddlers playing with LEGO” At first glance, the V4 image looks authentic.
Upon closer inspection, we can see that imaes looks more like . Children's faces look like rubber masks, their hands are a mess, and the LEGO pieces look like vague replicas. V5 resembles a real photo. The level of detail is remarkable.
Wider range of supported styles
The team behind Midjourney have touted V5 as having five times the stylistic range of its predecessor.
This means you can use longer and more descriptive prompts written in entire sentences to prompt a variety of visuals from landscapes to architecture.
The algorithm has also been enhanced for natural language processing, resulting in more accurate results.
In addition, many issues with V4 have been fixed such as improved handling of large groups of people, more realistic hands with the right number of fingers most of the time, and fewer image artifacts.
More “accurate” and predictable outcomes using image prompts
In contrast to its predecessors, Midjourney V5 offers a greater variety of images and is less restricted to specific artistic styles.
Holz suggests that to achieve the desired effect, it is necessary to use longer prompts that explicitly describe details in the image, such as lighting and mood.
Holz says the current v5 model is an alpha version that will undergo significant changes before its final release. The final version, like previous versions, will have a more beginner-friendly default style. Creatives shouldn't expect to see the current v5 style in the future.
Biomorphic skyscraper and snow-covered mountains:
MidJourney V5 Higher Resolution
MidJourney users have been hoping for an image resolution upgrade, and we’re excited to announce that the team is getting closer to making it happen.
By default, Midjourney V5 images are rendered at 1024 x 1024 pixels, up from 512 x 512 pixels in V4.
V4 images had to be upscaled to reach that resolution, so this is a significant improvement. In the future, we might see V5 upscalers that can increase the resolution to 2048 x 2048 pixels.
However, at the moment, 1024 x 1024 pixels is both the default and maximum size.
It's also quite a small resolution to use these images on a larger format. You can enlarge them with another tool that uses AI, like deep-image.ai.
Deep-image.ai allows you to automatically increase the resolution of the images by 4 times - up to 15,000 x 15,000 px.