Generative image AI platform Midjourney has introduced its V1 Video Model:
This Tweet is xex onlinecurrently unavailable. It might be loading or has been removed.
"It's fun, easy, and beautiful," the company posted on X on Wednesday. "Available at 10$/month, it's the first video model for everyoneand it's available now."
This appears to be a dig at competing generative-AI video programs. OpenAI's Sora is available for ChatGPT Plus and Pro users, for $20/month or $200/month, respectively, and Google's Flow is $249/month. (Adobe's Firefly starts at a compatible $9.99/month for up to 20 five-second videos, and Runway's Gen-4 Turbo video starts at $12/month).
You May Also Like
SEE ALSO: Google's AI Mode search tool gets a voice
Midjourney is still charging eight times more to produce a video than an image, and each job will produce four five-second videos. In a blog post on Midjourney's website, founder David Holz explained this and wrote that the prices will be hard to predict. The team will watch how V1 is used over the next month and adjust from there.
Holz called V1 a "stepping stone," as Midjourney ultimately wants to create "real-time open-world simulations." The building blocks to this, Holz wrote, is image models, video models of those images, 3D models, and doing this all quickly (real-time models). Midjourney plans on building these models individually and releasing them, with version one of its Video Model out now.
Midjourney users can create images in the platform as usual and now press "Animate" to make them move. They can choose to do this automatically or manually and choose between low motion (for more ambient scenes) and high motion. Videos can be "extended" around four seconds at a time, four times in total.
Users can animate images from outside of Midjourney, as well. For now, V1 is web-only.
"We ask that you please use these technologies responsibly," Holz wrote.
As Mashable's Timothy Beck Werth reported for the launch of Google's Veo 3, misinformation experts have sounded the alarm that AI video may soon be indistinguishable from real video. (The recent viral emotional support kangaroo video shows that that's already happening.) AI generation has been used by bad actors, such as to create explicit deepfakes (which is now a federal crime in the U.S.)
Related Stories
- Asana bug in new AI feature may have exposed data to other users for weeks
- The AI-generated ad that aired during the NBA finals took $2,000 and 2 days to make
- YouTube very quietly loosened its content moderation rules
- Meta reportedly replacing human risk assessors with AI
- Meta's AI tool Llama 'almost entirely' memorized Harry Potter book, study finds
"Properly utilized it's not just fun, it can also be really useful, or even profound — to make old and new worlds suddenly alive," Holz continued.
V1 is launched amid a recent lawsuit Disney and Universal filed against Midjourney. The suit claims that the platform illegally trained on copyrighted content and called Midjourney a "bottomless pit of plagiarism."
Disclosure: Ziff Davis, Mashable’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.