Midjourney has released V7 Alpha, its first new AI image generation model in nearly a year, featuring voice prompting capabilities and a faster draft mode. The launch comes a week after OpenAI debuted a new image generator in ChatGPT that quickly gained popularity.
According to Midjourney CEO David Holz, V7 represents a “totally different architecture” that is “much smarter with text prompts” and delivers higher image quality with “beautiful textures.” The model is said to produce better coherence in details for bodies, hands, and various objects.
Key features of Midjourney V7 Alpha
- Voice prompting: Users can speak directly to the model through the alpha website, allowing for a more natural workflow
- Draft Mode: Renders images 10 times faster at half the cost of standard mode, though at lower initial quality
- Personalization requirement: Users must rate around 200 images to build a personalization profile before using V7
- Two operational modes: Turbo (higher performance but costlier) and Relax
The new voice input feature works alongside Draft Mode, enabling users to enter what Midjourney describes as a “flow state” of creative drafting. Users can speak instructions like “make this more detailed” or “make this darker” to adjust generations in real-time.
Several standard Midjourney features aren’t yet available for V7, including image upscaling and retexturing. According to Holz, these will arrive “in the near future, possibly within two months.”
Mixed reception from users
Unlike previous Midjourney releases that received widespread praise, initial reactions to V7 have been mixed. Some users have expressed disappointment, with AI influencer Ethan Mollick noting that “the problem with the new V7 is that V6 was already really good.”
Critics point out issues with prompt adherence, anatomical accuracy (particularly hands), and text generation capabilities. Some users have reported that “identical prompts from V6 are worse in V7.”
However, others have praised V7’s image quality, with AI artist Tatiana Tsiguleva describing it as a “huge jump in quality.”
Midjourney, a San Francisco-based company founded in 2022 by David Holz, has not raised outside funding but was reportedly expecting around $200 million in revenue in late 2023. The company faces several lawsuits alleging copyright infringement for training its AI tools on web-scraped images without creators’ consent.
Quellen: TechCrunch, VentureBeat