Midjourney has released the alpha version of its V7 model, which boasts an entirely new AI image generation model. This latest iteration is reportedly much more intelligent in processing text prompts and produces higher-quality images with improved textures, bodies, and hands. Notably, AI image generators often struggle to accurately depict hands, but V7 has demonstrated the ability to generate photorealistic images of human hands, as evident from photos shared by users on social media.

The V7 model introduces a new feature called “Draft Mode,” which allows for faster image rendering at half the usual speed, albeit with rougher and less detailed results. This mode is priced at half the cost of a standard generation and is ideal for iterating on ideas, particularly during collaborations or when exploring different creative directions. Draft Mode also supports voice input, eliminating the need for typing. If a generated image in Draft Mode is satisfactory, users can opt to “enhance” or “vary” it to re-render it at full quality.

Initially, V7 will offer Turbo and Relax modes for standard rendering, with the Turbo mode costing twice as much as a normal speed job on the V6 model. Although Midjourney needs more time to optimize its standard speed mode for V7, it plans to make this option available in the future. Currently, the new mode lacks certain capabilities, including upscaling, editing, and retexturing, which will default to the V6 model. Midjourney promises to roll out updates for the model every week or two over the next two months.

To access the alpha version of V7, users must first unlock their personalization profile. Midjourney describes personalization as a “style assistant for your image creations,” as it teaches the AI to recognize individual visual preferences. Users must rank at least 200 pairs of images to create a V7 Global Personalization Profile and test the model. While personalization is enabled by default for V7, users can choose to disable it if desired.