Midjourney has officially launched the alpha version of its V7 image generation model, introducing significant advancements in performance, usability, and creativity. The AI community is now invited to test and explore the latest capabilities.
V7 offers noticeable improvements in image quality, prompt comprehension, and overall coherence. According to Midjourney, the model is especially adept at interpreting complex text prompts, generating visually accurate and aesthetically rich images. Enhancements include more realistic textures and improved rendering of fine details such as hands, objects, and body proportions.
One of the most notable additions is the model’s personalisation feature, which is enabled by default once a five-minute setup process is completed. This feature allows users to fine-tune outputs based on their preferences and can be toggled on or off at any time. Midjourney sees this as a breakthrough in aligning AI output with user intent.
Draft Mode: Faster and Cheaper Image Creation
Accompanying the V7 model is a new Draft Mode, which generates images ten times faster and at half the cost of standard outputs. This mode facilitates rapid prototyping of ideas and enhances the creative workflow.
Draft Mode also introduces conversational capabilities. Through Midjourney’s web interface, users can request modifications to images—such as changing a cat to an owl or transitioning from day to night—and the system will automatically adjust prompts and regenerate images accordingly.
Adding to its innovative edge, Draft Mode supports voice input. By pressing a microphone button, users can speak their image ideas aloud and watch as images are rendered in near real-time. This interaction provides a fluid and intuitive creative process.
While draft images are lower in resolution, they retain the same behavioral and aesthetic traits as those generated in standard mode. If a user is satisfied with a draft, it can be enhanced or varied for full-quality output.
Speed Tiers and Cost Structure
V7 will launch with two operational speed modes: Turbo and Relax. Turbo jobs are priced at double the standard rate, while Draft Mode jobs cost half. The traditional standard mode is under optimization and will be made available soon.
Feature Roadmap and Compatibility
Midjourney confirmed that certain features—such as upscaling, editing, and retexturing—will temporarily fall back on the V6 model. Support for mood boards and Style Reference (SREF) is already functional, with ongoing performance improvements expected.
A key upcoming feature is character and object referencing in V7, designed to further enhance consistency in storytelling and design applications.
The development team plans to roll out new features every 1–2 weeks over the next 60 days. Midjourney emphasized that V7 is an entirely new architecture and may require users to adapt their prompting styles.
With Midjourney V7, users are getting a more intelligent, flexible, and user-friendly image generation tool. Its combination of speed, customisation, and voice interaction positions it as a game-changer for designers, artists, marketers, and AI enthusiasts. As the alpha evolves, community feedback will be crucial in shaping the next phase of creative AI tools.
