Midjourney has launched Version 7 (V7), its first new AI image generation model in nearly a year. The highly anticipated update introduces a completely new architecture, smarter text prompt handling, enhanced image coherence, and built-in personalization.
Rolling out in alpha since Thursday, V7 arrives shortly after OpenAI’s viral image generator update in ChatGPT, but Midjourney’s V7 stands out with its unique approach and performance upgrades.
What’s New in Midjourney V7?
- Personalization by default: Users are required to rate ~200 images to activate a profile that aligns image generation with their style preferences.
- Better image quality: Enhanced rendering of textures, hands, faces, and object details.
- Improved text prompt understanding: Smarter interpretation of descriptions for more accurate outputs.
- Draft Mode: Generates images 10x faster and at half the cost, though with lower quality (upgradeable with one click).
- Two Modes: Choose between Turbo (faster, more expensive) and Relax.
Features Still to Come
While upscaling and retexturing aren’t yet available for V7, Midjourney CEO David Holz mentioned they are expected within the next two months.
Holz also emphasized that V7 may require different prompting techniques, encouraging users to explore and experiment with the model’s capabilities.
Midjourney’s Growth and Vision
Founded in 2022 without external funding, Midjourney has quickly grown into a major player in the AI art space. The company reportedly generated $200 million in revenue in 2023 and is now assembling a hardware team and training models for video and 3D object generation.
V7 reflects Midjourney’s continued innovation in AI visuals, providing artists and creators with a powerful, flexible tool that evolves with their creativity.
