Modern tech office interior, wide angle shot, natural lighting through floor to ceiling windows, multiple computer screens displaying colorful AI generated images, 85mm lens, f2.8, soft depth of field, Canon R5, cinematic color grading.
Created using Ideogram 2.0 Turbo with the prompt, "Modern tech office interior, wide angle shot, natural lighting through floor to ceiling windows, multiple computer screens displaying colorful AI generated images, 85mm lens, f2.8, soft depth of field, Canon R5, cinematic color grading."

Midjourney World Models and V7: What You Need to Know

Midjourney just released their office hours notes, and I want to break down the most important updates. The team has shifted their focus to world models, with data gathering starting this December and training kicking off in January. We should see the first features from this development by February.

The biggest news? Video capabilities and V7 models are dropping in January. The team plans to test these with guides and moderators throughout December, making sure everything works properly before the public release.

I’m particularly interested in the new retexture mode they announced. It gives you much tighter control over image editing, which pairs well with their recent external image editing capabilities. If you’ve been following my coverage of AI image editing tools, you’ll know this puts Midjourney in direct competition with specialized editors like Flux.1 (https://adam.holter.com/flux-1-tools-ai-image-editing-gets-surgical-precision/).

They’re also rolling out mood boards and personalization profiles soon. This means you’ll be able to create multiple profiles and drop in multiple reference images to customize your results. Think of it as having different presets for different types of projects.

The team specifically mentioned they’re slowing down during the holidays.

My take? Midjourney is positioning itself as more than just an image generator. With world models, video capabilities, and advanced editing tools, they’re building an entire creative platform. The focus on world models suggests they’re aiming for more coherent, contextual image generation – something that’s been a weak point for most AI image tools.

I’ll update you as soon as these features start rolling out. For now, if you want to stay current with major AI developments, check out my analysis of other recent releases like OpenAI’s Sora (https://adam.holter.com/openai-sora-leak-artists-fight-back-against-tech-giant/) and Runway Frames (https://adam.holter.com/runway-frames-their-new-ai-image-model-makes-photos-you-cannot-tell-apart-from-reality/).