There is a lot of debate around artificial intelligences that are sold as open source, such as Twitter’s xAI (now X). The truth is that we don’t care about that debate and even more so now that Mochi 1 has just been announced, the first open source AI to generate videos and completely free that looks spectacular.
Currently, as far as we know, there are several AI systems that focus on video creation. Most of these solutions are of quite poor quality and have a huge amount of glitches in the background, transitions, movements and so on.
The great promise is Sora, the OpenAI solution that is currently announced, but is still not available. While this AI for video generation arrives, we have something quite interesting already available such as Mochi 1, which we will talk to you about in more detail.
Mochi 1, the AI ​​for Open Source videos, is finally available
As we have mentioned, there are many models that allow you to generate video, but none of them work very well. Some even generate truly terrifying shorts far from what one has requested through a prompt.
One of the strong points of Mochi 1, which belongs to Genmo AI, is precisely that it is open source under the Apache 2.0 license. It represents a very important point when it comes to the democratization of video technology with AI.
What stands out about this solution is its architecture with more than 10,000 million parameters. We are talking about the largest video generation model that exists in open source. The aim, as stated by Para Jain, CEO of Genmo AI, is that they seek to ensure that any user can become a creator.
Grade
@Gradio
Mochi 1
Dramatically closes the gap between closed and open video generation models. ✅
Apache 2.0 license 🤯
High-fidelity videos
Strong prompt adherence
Model available on 🤗 Hub https://t.co/XAN6N8AHY2October 22, 2024 • 19:41
730
12
Mochi 1 is characterized by offering realistic movement that respects the laws of physics in detail. This, precisely, is one of the weakest points of the solutions currently available.
Another strength of this technology is that it follows the instructions given in the command entered. It also offers us a greater level of detail of the characters, settings and actions. Always seeking to offer a fluid and pleasant experience that makes it difficult to identify that it has been generated by an AI.
It also has another great strength and that is that the easy features are coherent and fluid. Many times we find in the videos, crazy transitions, deformed faces and other aberrations that seem straight out of a terrifying nightmare.
This solution can be used for free, although with the limitation of creating up to 4 videos every 6 hours up to a maximum of 30 videos per month. All content will come with Genmo watermark and will be for personal use. They offer a Lite license for $8 per month that allows you to generate up to 8 videos every six hours with a maximum of 80 videos per month. There is no watermark here and they allow commercial use with rights monetization.
Finally, a category called Standard is offered that costs $24 per month. This variant allows you to generate up to 32 videos every 6 hours with a maximum of 180 videos per month. It allows you to generate video without watermark with hidden mode for greater privacy, the creator has the rights to the video and priority access is offered to new models.
