Big Tech enters his AI video -era. Here are the main models

Sora joined the Chatgpt family at the end of 2024. It is a fairly user-friendly program, but it does not have the same conversation user interface as Dall-E 3-Sie cannot “chat” with Sora to carry out follow-up revisions. Instead, it is more reminiscent of traditional AI creative services. You have a panel to adapt the dimensions, length and the stylistic feeling of your video. You can enter an input request or upload an image for Sora for Animate and use some editing options to perfect your video from there. Sora videos also automatically come with watermarks and refer to their AI jumps.

Sora is only available for paying chatt users. If you are a chatt plus user (20 US dollars per month), you will receive 50 credits for priority generations per month, with videos up to 5 seconds at 720p. The upgrade to the pro level (200 US dollars per month) receives higher monthly credits, including 500 videos, which were created with a relaxed generation with a relaxed generation and unlimited videos. Pro subscribers can also create videos with a higher resolution on a maximum of 1080p, expand the duration of their videos to 20 seconds and have the opportunity to download videos without a watermark.

Openaai's data protection guideline says that it can train on its content unless they refuse. To do this in Sora, go to settings> General and switch off Improve the model for everyone. You can also exclude your videos from public exploration feeds in settings.

Firefly's independent AI -Videogenerator came from a private beta in February and is now available for you. If you are familiar with Firefly's AI image tools, the video technician will feel familiar. You can adapt your clip with the left side. You have the additional advantage that you can select the desired type of movement (zoom in to the right and left, etc.). You can also select the desired camera hinge if you want to imitate drone material with an air view.

Some creative cloud plans include Firefly Access, including if you pay for a single program or all Adobe apps. Here you can check and compare options. If you do not have an existing Adobe plan, you can try out the Firefly Standard plan ($ 10 per month) to create up to 20 videos per month. If you need more generation loans, you will receive up to 70 videos per month with the Pro plan ($ 30 per month). Both Firefly plans are delivered with unlimited AI image generation. Your Firefly videos are at 1080p without audio for 5 seconds.

According to Adobe, videos that were created with Firefly are commercially secure, and its AI guidelines say that it is not trained on their content. Firefly videos have no visible watermark, but the content registration information is automatically connected to their work. Firefly models are trained on licensed and public domestic content.

The Deepmind Research Lab from Google has been working on the company's AI video model for some time. Google announced the project at its developer conference last year and shows some of the possible applications in a clip with Donald Glover. In December 2024, Google announced that VEO 2 for approved users on Vertex AI, the machine learning platform from Google Cloud, was brought into a private preview. You can register for the waiting list here.

According to what we have seen so far, VEO can process text-to-video and image-to-video input requests. We don't know how long the clips will take or whether you have audio. According to Google, videos are created on a maximum of 1080p and the synthidal watermark of Google is embedded. Google Cloud's general AI guideline indicates that you do not train on customer data without permission. Google did not answer inquiries about further information on the development and publication of VEO.

If you don't want to wait, you can get an impression of what Veo 2 can do on YouTube. Last month, the company announced that youtube shorts manufacturers could now create a new function of VEO with AI-generated backgrounds for their videos. In my first tests, the creation process was smooth, but I reserve the judgment until the program is fully shaped and available for everyone.

AI enthusiasts could recognize the runway as the start-up that has put together the popular AI image generator stable diffusion. You may also recognize Runway from a pioneering deal that it did with a big film studio last autumn. Lionsgate agreed to open his catalog -thousands of hours of films such as The Hunger Games and John Wick and TV shows such as Mad Men -to create custom AI models for the studio. Although Runway is technically not one of the large technology companies, the AI ​​video tools are worth mentioning.

During my short examination of the service, I was impressed by the prompt building tools and the general lightness to deal with. I also used the service as part of the Magic Media app from Canva, which is comfortable when you are a canva lover. You can use Runway for free in the web app with 125 monthly credits – you will use about 20 credits with each generation, so this is a rather low limit. The upgrade ($ 15 per month or $ 144 per year) gives you 625 monthly credits, access to newer models and the possibility to increase videos to 4K and download without a watermark.

According to Runway's terms of use, you can train your AI for your input requests and the resulting videos, but does not retain ownership of you. The data protection guideline also states that Runway can give in to your information to partners, business partners and marketing. The videos you make are automatically private.

Last but not least is the AI ​​video model from Meta, Film Gen. The project is currently only a research concept -it is not publicly available without it being a word about when it may be. Maybe like with Google and YouTube, we will first see on his social platforms, Instagram and Facebook, first in AI-powered functions. (We already have a number of other AI features that eat up space in our feeds.)

Thanks to a published research paper, we know that films can be 1080p HD and for up to 16 seconds with 16 frames per second. The most remarkable thing that is used for the films is the possibility of a synchronized audio. Meta said that film genes could also be used to create sound effects, ambient noise and instrumental music for up to 45 seconds. There is always a chance that this function does not make it to the final cut, but it would give meta a great advantage over all of its competitors.

The AI ​​models from Meta for Chatbot and Image Generator are on publicly available content from Facebook and Instagram (European users can unsubscribe) and licensed information are trained.