Midjourney V6 version has finally arrived!

After months of anticipation, the awaited release of the new Midjourney V6 version has finally arrived. David Holz shared the exciting news today on the Discord server:

Hello @everyone! We're thrilled to announce that the alpha version of our V6 model is now available for community testing throughout the winter break, starting tonight. To activate it, simply select V6 from the dropdown menu under /settings or type --v 6 after your prompt. What's noteworthy about the V6 base model.

1. Significantly improved prompt following and support for longer prompts.

2. Enhanced coherence and expanded model knowledge.

3. Improved image prompting and remix capabilities.

4. Minor text drawing ability (text must be written in "quotations" and --style raw or lower --stylize values may enhance the effect).

5. Try /imagine a photo of the text "Hello World!" written with a marker on a sticky note --ar 16:9 --v 6.

6. Improved upscalers with both 'subtle' and 'creative' modes (doubles resolution).

(Buttons for these modes can be found under your images after clicking U1/U2/U3/U4).

At launch, the following features/arguments are supported:

--ar, --chaos, --weird, --tile, --stylize, --style raw, Vary (subtle), Vary (strong), Remix, /blend, /describe (only the v5 version). Features not yet supported but expected in the coming month: Pan, Zoom, Vary (region), /tune, /describe (a new v6 version).

Styling and prompting for V6 require a different approach. Here are some key points:

1. V6 is more sensitive to prompts; explicit instructions are recommended.

2. Avoid vague terms like "award-winning," "photorealistic," "4k," and "8k."

3. For a more photographic and less opinionated result, consider using --style raw.

4. Lower values of --stylize (default 100) may enhance prompt understanding, while higher values (up to 1000) may improve aesthetics.

Community collaboration is encouraged in the prompt-chat to share insights on V6 usage. However, please note:

- This is an alpha test; frequent changes may occur without notice.

- The exact model may not remain the same in the future as V6 progresses toward full release.

- Speed, image quality, coherence, prompt following, and text accuracy are expected to improve over the next few weeks.

V6 is slower and more resource-intensive than V5 initially but will become faster with optimization. Relax mode is supported (about 1 GPU/min per imagine and 2 GPU/min per upscale).

In terms of community standards:

- The model can generate highly realistic imagery, and moderation will be strict to enforce community standards.

- Avoid creating images that may cause drama or being disrespectful.

Additional details about V6:

- V6 is the third model developed from scratch on our AI superclusters, with a 9-month development period.

- While V6 is not the final step, it represents a profound progression intertwined with the collective power of our imaginations.

Share your V6-generated images in v6-showcase or provide feedback in ideas-and-features. Wishing you joy, wonder, responsibility, and respect in utilizing these incredible powers. Have fun and happy holidays!

#midjourney #ai