Midjourney V7 Update: Latest Details on What's Coming

Get the scoop on the newest developments straight from the Midjourney team. The recent office hours shared some interesting details about the upcoming version 7, including testing progress, new features, and what to expect upon release.

If you use Midjourney often, keeping up with updates can help you make the most of your image generation. These changes aim to improve how the AI understands what you want and how fast it works.

Midjourney Version 7 is Getting Closer

The big news is that Midjourney Version 7 (V7) is under development. The team has started testing it with a small group of guides and moderators. So far, they are mostly happy with the results.

Like any testing phase, there are a few small issues they are working through. One example mentioned was the AI sometimes splitting images in two, especially with tall subjects in a square frame. These are the kinds of things they fix during testing.

Omni Reference Training Starts

A feature called Omni reference wasn't in the first test build of V7. This is a special step in the AI's training process. They are starting the Omni reference training now. This part can take between 1 to 3 days to finish. It's something that happens after the main training is done.

Faster Generations

Speed is another area they are working on for V7. The goal is to make image generations faster. They are aiming for around 15 seconds in standard mode and potentially 5 seconds in turbo mode. However, they are not certain yet if they will hit those exact targets when it launches. Faster creation means you can iterate quicker and try out more ideas.

Speaking of creating more ideas quickly, have you explored tools that can help manage your Midjourney workflow? The Midjourney Automation Suite from TitanXT can help streamline your process, letting you focus on bringing your visions to life.

What to Expect When V7 Launches

When V7 is ready for the public, it will likely release in an alpha stage first. This means it won't be turned on by default for everyone. Users will have the option to switch to V7 alpha to test it out. This allows the team to gather wider feedback before making it the default version.

They are also working on updates to the user interface. This could mean changes to how you interact with Midjourney, possibly outside of Discord.

Performance Insights

Here’s how V7 is performing according to the team:

• Image Quality: The image quality and how well parts of the image fit together look good. The team is pleased with this.

• Prompt Understanding: They rated the AI's ability to understand prompts as a 7 out of 10 compared to V6. This means that for every 10 things that didn't work in V6, about 7 of them should work in V7. This is a good improvement.

Areas Needing Work

There are still a couple of areas that need improvement:

• Text: V7 is currently not good at handling text within images. This is a known challenge with AI image models.

• Variety: The variety in the generated images needs to be adjusted. They are working on tweaking this.

Other Upcoming Updates

Besides V7, there are other updates in the works.

Editor Updates

An update to the Midjourney editor is expected to be released soon, potentially next week. This update might include features related to layers, retexturing parts of an image, and possibly combining different editor functions. There was also a mention this could potentially involve Discord, but that part wasn't confirmed.

Managing layers and editing can add steps to your workflow. Tools like the Midjourney Automation Suite from TitanXT are designed to help handle repetitive tasks, giving you more time to focus on the creative edits and refinements.

Video Model Training

Training for a video model is scheduled to begin in the first week of April. This suggests that video generation capabilities could be a future addition to Midjourney.

Exploring new AI capabilities like video takes time and resources. Automating parts of your image generation process with the Midjourney Automation Suite frees up valuable time you can use to experiment with upcoming features like video or refine your current image prompts.

Summary

That’s the latest from the Midjourney office hours. V7 is progressing well in internal testing with improvements in prompt understanding and image quality, though text and variety are being refined. Omni reference training is underway, and speed goals are ambitious. Expect V7 to launch in alpha first. Editor updates are coming soon, and video model training starts in April. Stay tuned for more news as these updates get closer to release.