Midjourney has launched the alpha version of its V7 image generation model, introducing significant upgrades aimed at enhancing user experience and image quality. This new model boasts improved text prompt comprehension, creativity, and coherence in generating images. According to Midjourney, V7 excels in understanding user instructions and offers exceptional texture detail and accuracy in rendering bodies, hands, and various objects.
One noteworthy feature of V7 is the activation of model personalization, which users can unlock in about five minutes. This personalization allows the AI to better interpret individual user preferences and artistic intents, elevating the overall creative process. Users can toggle this feature on or off, making it adaptable to various needs.
Complementing V7 is the ‘Draft Mode,’ designed to produce images up to ten times faster and at half the cost. This efficiency has facilitated the introduction of a novel “conversational mode” within the web interface, enabling users to request modifications verbally. For example, users can change the subject of an image and the AI will promptly update the output based on the new instructions. Additionally, the Draft Mode allows for voice input, letting users express their ideas verbally while generating images in near real-time.
While draft images may not match the quality of those produced in standard mode, they retain consistent behavior and aesthetic characteristics. The V7 model offers two speed modes—Turbo and Relax—though Turbo jobs will come at double the cost of standard jobs, while draft jobs will be more affordable.
Midjourney is already planning a robust schedule for releasing updates and new features over the next two months, including improved capabilities for mood boards and a new reference tool for characters and objects. As V7 is designed with its own set of strengths and weaknesses, Midjourney encourages users to experiment and share feedback for further enhancements.
