First came the rise of AI chatbots, then image generators blew up. Now, tech companies are rushing to release AI video generators.
During the past year, nearly every major tech company has announced some kind of AI video model they’ve been cooking. Each company has its own timeline, which can make it hard to keep up with who’s done what. To save you from searching, I’ve run down every major AI video program and compiled my early insights on the testing I’ve done on the programs available now. They aren’t all built the same, and there are noticeable difference even across one company’s AI products. For example, I’ve seen some of my favorite image generator features pop up in the video models, while others are noticeably absent.
AI videos are a huge leap forward in a company’s AI creative offerings, and they’re something worth keeping an eye on as generative AI become a bigger part of the content we create and see online. This is especially true as the advancement of the tech comes at a time when the legality, ethics and other concerns continue to surround AI creative offerings.
This is everything you need to know about the major AI video generators. This list is regularly updated with the most recent info on each generator. For more, check out the best AI image generators.
What are AI video generators?
AI video generators are one of the latest ways tech companies are using generative AI. These programs use text-to-video and image-to-video technology that lets you create short video clips. You enter a short description called a prompt, or upload an image to animate, and the software creates a clip entirely made with gen AI. These AI videos tend to be between five and 10 seconds long, and only Google’s Veo 3 has synchronized audio. Because this tech is new, errors — called hallucinations — are possible.
What AI video models I can use right now?
Some examples of AI video generators you can use now are Sora by OpenAI, Veo 3 by Google and Adobe Firefly. They are all are paid programs that produce decent results and let you customize your shot with control panels. Runway, an AI start-up that co-created the Stable Diffusion image generator, is another AI video option with freemium plans. Other AI start-ups like Luma, Pika and Ideogram are also available.
CNET, Lily Hailyeh
Sora joined the ChatGPT family at the end of 2024. It’s a pretty user-friendly program, but it doesn’t have the same conversational UI as Dall-E 3 — you can’t “chat” with Sora to make follow-up revisions. Instead, it’s more reminiscent of traditional AI creative services. (Disclosure: Ziff Davis, CNET’s parent company, in April filed a lawsuit against OpenAI, alleging it infringed Ziff Davis copyrights in training and operating its AI systems.)
In Sora, you have a panel to customize your video’s dimensions, length and stylistic feel. You can enter a prompt or upload an image for Sora to animate, and you can use a few editing options to perfect your video from there. Sora videos also come automatically watermarked, designating their AI origins.
Sora is only available to paying ChatGPT users. If you’re a ChatGPT Plus user ($20 per month) you’ll get 50 priority generation credits per month, with videos up to 5 seconds long at 720p. Upgrading to the Pro tier ($200 per month) gets higher monthly credits, including 500 videos created with priority/fast generation and unlimited videos with relaxed generation. Pro subscribers can also create higher-resolution videos at a max of 1080p, extend the duration of their videos up to 20 seconds and have the option to download videos without the watermark.
OpenAI’s privacy policy states that it may train on your content unless you opt out. To do that in Sora, go to Settings > General, and turn off Improve model for everyone. You can also exclude your videos from public explore feeds in settings.
Screenshot/James Martin/CNET
Google’s Veo 3 AI video generator enthralled AI enthusiasts when it was publicly released at its 2025 I/O developers conference. In my testing, I found Veo 3 to be a nice step up from my disappointing experience with Veo 2, but the true difference is audible: Veo 3 can generate and automatically sync audio to your video clips.
Veo 3 videos are eight seconds long, at 720p resolution and are automatically generated in a horizontal 16:9 format. It’s a very capable program and the videos are detailed and fun, but there are some less exciting notable aspects to it. Veo 3 is only available through the pricey Gemini Ultra plan, and it doesn’t have editing tools, which is a bummer. Future updates could make Veo 3 more widely accessible and advanced.
Videos automatically have Google’s SynthID watermark embedded (not visible on the clips). Google Cloud’s general AI policy states it does not train on customer data without permission.
You can also give Veo a test run on YouTube. Creators can now create AI-generated backgrounds for their Shorts videos, thanks to a new feature powered by Veo.
Midjourney
Midjourney is one of the most popular AI image generators, and it just released its first AI video model called V1. You can use Midjourney to create video clips between 5-and-21 seconds long in 720p resolution. You can use Midjourney through Discord or its website. Right now, video generation is paywalled for users, but it’s one of the cheaper options at $10 per month.
Midjourney’s privacy policy says it can use personal information and information included in your prompts to improve its service. If you create in Stealth mode, then your AI images will be private; otherwise they will be shared in a public gallery.
Adobe
Firefly’s standalone AI video generator is available for you to use now, on your computer and through its Firefly mobile app. If you’re familiar with Firefly’s AI image tools, the video generator set up will feel familiar. The left-side panel lets you customize your clip, and it has the added benefit of letting you select the kind of motion you want (zoom in and out, move right and left, etc). You can also select the camera angle you want, like if you wanted to mimic drone footage with an aerial view.
Some Creative Cloud plans include Firefly access, including if you’re paying for a single program or all of the Adobe apps. You can check and compare options here. If you don’t have an existing Adobe plan, you can try the Firefly standard plan ($10 per month) to create up to 20 videos a month. If you need more generation credits, the Pro plan ($30 per month) gets you up to 70 videos a month. Both Firefly plans come with unlimited AI image generation. Your Firefly videos will be 5 seconds long, at 1080p with no audio.
Adobe says that videos created with Firefly are commercially safe, and its AI policy states it will not train on your content. Firefly videos don’t have a visible watermark, but its content credentials are automatically attached to your work. Firefly models are trained on licensed and public domain content.
Katelyn Chedraoui/CNET
AI enthusiasts might recognize Runway as the start-up that co-created the popular AI image generator Stable Diffusion. You might also recognize Runway from a landmark deal it made with a major film studio last fall. Lionsgate agreed to open up its catalog — thousands of hours of movies like The Hunger Games and John Wick and TV shows like Mad Men — to be used to create custom AI models for the studio to use.
During my brief testing of the service, I was impressed with the prompt-building tools and the general ease of finding my way around. I’ve also used the service before as part of Canva’s Magic Media app, which is convenient if you’re a Canva lover. You can use Runway for free on its web app, with 125 monthly credits — you’ll use about 20 credits with each generation, so it’s a pretty low limit. Upgrading ($15 per month or $144 annually) gets you 625 monthly credits, access to newer models and the ability to upscale videos to 4K and download without watermarks.
Runway’s terms of service says it can train its AI on your prompts and the resulting videos but doesn’t retain ownership over them. Its privacy policy also states that Runway may disclose your information to affiliates, business and marketing partners. The videos you make are automatically private.
What are some other AI video projects?
Notably absent from this list is Meta. The company has devoted its resources to develop AI, but it doesn’t have a publicly available AI video generator. It teased a version of one in October 2024. Here’s what we know so far.
Meta/Screenshot by CNET
Meta’s AI video model — Movie Gen — is only a research concept right now and not publicly available, with no word on when it may be.
Thanks to a research paper Meta published, we know Movie Gen videos could be 1080p HD and up to 16 seconds long at 16 frames per second. The most notable thing going for Movie Gen is the possibility of synchronized audio. Meta said that Movie Gen could also be used to create sound effects, ambient noise and instrumental music up to 45 seconds long. There’s always a chance this feature doesn’t make it to the final cut, but it would give Meta an edge.
Perhaps like with Google and YouTube, we’ll see some AI-powered features pop up first on its social platforms, Instagram and Facebook. (We already have a number of other AI features eating up space on our feeds.) Meta’s AI models for its chatbot and image generator are trained on publicly available Facebook and Instagram content, as well as licensed data.
For more, check out our guide to writing the best AI image prompts and the best AI chatbots.
