🚀 Midjourney Office Hours Highlights: March 5th

Not sure I want to publish any possible release dats for V7 anymore, however in terms of transparency, the team shared that internal testing is underway, and if all goes smoothly, the public release could be just two weeks away. Expect faster performance, enhanced upscaling capabilities, and the introduction of video features. We're also working on cloud infrastructure upgrades to boost platform speed. Community feedback will be crucial as we refine these updates, so stay tuned for opportunities to share your thoughts.

For a quick overview, head straight to the TL;DR; section otherwise keep on reading.

Extended Recap

🗓️ Version 7 Release Timeline

The much-anticipated Version 7 is on the horizon, with internal full-pipeline testing already underway. By the end of this week, guides and mods will begin testing an incomplete yet functional model. The final touches, including feature integration and speed optimization, are expected to wrap up next week. If all goes according to plan, the public release of V7 could be just two weeks away. However, potential delays may arise due to untested internal integrations and recent code refactoring. Initially, V7 will be available in alpha mode, meaning updates and changes could occur without prior notice.

⚡ Performance and Optimization

V7 promises to be a game-changer in terms of speed, with the default mode expected to be twice as fast as its predecessors. While a slower, higher-quality variant is under consideration, users might soon have the ability to adjust the speed-quality balance through sliders or controls. Post-release, the team plans to retrain the model for immediate quality improvements and is considering integrating new datasets for further enhancements.

🔍 Upscaler Improvements

The development of faster upscalers is in progress, with the potential for higher-resolution capabilities. The team is aiming to reintroduce a 4x upscaler, although significant behavior differences are anticipated, and the exact outcomes remain uncertain at this stage.

🔗 Omni Reference and Personalization Codes

The omni reference feature is set to be integrated during the final distillation phase, promising substantial improvements over Version 6. There's an ongoing debate about whether to break or maintain compatibility with existing personalization codes. While the current priority is to maintain backward compatibility, no absolute guarantees can be made.

🎥 Video Feature Development

Midjourney is preparing for extensive video training runs, set to commence within one to two weeks. The initial model release will focus on optimizing the balance between cost and quality, with plans for a subsequent Version 2 video model featuring significant improvements. Three workflows are under consideration: Text-to-Video, Image-to-Video, and Vari-to-Video, each with its own set of challenges and potential. The video model aims to match the speed of current image models, with aggressive cost reduction targets. Over the next six months, users can anticipate substantial improvements in video quality and cost reductions.

📽️ Video and 3D Feature Roadmap

Following the release of V7 and the initial video model, additional video modes are planned. However, 3D features have been delayed and will be prioritized after the completion of these milestones.

☁️ Cloud Infrastructure and Server Restructuring

Midjourney is undergoing significant cloud restructuring efforts to enhance service performance. Over the next six months, server migration will aim to improve platform speed and potentially reduce costs. These infrastructure changes are logistically complex and resource-intensive.

🛠️ Editor and Batch Mode Improvements

The development of editor enhancements continues, alongside progress in large-scale batch modes and workflows.

🤝 Community Participation and Feedback

Midjourney remains committed to transparency with its community. Rating parties for aesthetic and feature evaluations are planned post-V7 release, with a major rating event specifically for upcoming video models. The team values community feedback and aims to keep users informed about internal processes and feature progress.

TL;DR

🗓️ V7 Release: Testing underway, public release in ~2 weeks, potential delays.

⚡ Performance: V7 twice as fast, speed-quality balance controls post-release.

🔍 Upscalers: Faster, higher-resolution upscalers in development.

 🔗 Personalization: Omni reference improvements, compatibility debates ongoing.

🎥 Video Features: Extensive training, initial model release soon, three workflows.

📽️ Roadmap: Additional video modes post-V7, 3D features delayed.

☁️ Infrastructure: Cloud restructuring for better performance, server migration planned.

🛠️ Improvements: Editor and batch mode enhancements in progress.

🤝 Community: Rating events planned, commitment to transparency.

Stay tuned for more updates as Midjourney continues to innovate and enhance its platform!

If you want to support me, feel free to buy me a coffee ☕️ 

If you’re not subscribed yet to my newsletter Imagine Weekly, I’d be thrilled to welcome you on board!