OpenAI's Sora is one of the most impressive AI tools I've seen in my years covering the tech, but so far it's only accessible to a handful of professional creators and I don't see it becoming widely available anytime soon.
We've seen a ton of memorable videos, from documentaries about astronauts to music videos about watching the rain. We've also seen a short film about a man with a balloon head and a commercial for Toys R Us.
OpenAI Chief Technology Officer Mira Muratti initially suggested Sora would be available sometime this year, but that has now apparently been delayed, with the latest update suggesting that OpenAI developers are having trouble making Sora easy to use.
The company said it's now focused on both safety and ease of use, which will likely include ensuring guardrails prevent the robots from replicating real people or being used for misinformation campaigns.
Sora alternatives already available
While you wait for Sora, there are already some great AI video tools that can help you create a variety of clips, styles, and content. Try these tools: Alternatives to Sora include Pika Labs and Runway, Luma Labs Dream Machine, and more.
Sora's biggest selling points were its more natural movements and longer initial clip generation times, but with the arrival of the Dream Machine and Runway Gen-3, some of those unique features have already been replicated.
Currently, there are two categories of AI video models, which I divide into first and second generation: The first generation are the well-established video diffusion-based models, including models such as Runway Gen-2, Pika 1.0, Haiper, as well as models found in Leonardo and NightCafe.
The main limitation of this early generation of AI video tools is duration: most can't deliver consistent movement for more than 3-6 seconds, and some struggle to go beyond 3 seconds. This, plus the small context window, makes consistency difficult.
However, second-generation models like Runway's Gen-3 and Luma Labs Dream Machine (as well as unavailable models like Kling and Sora) have a much longer initial generation, better motion understanding, and improved realism.
Runway
Runway is one of the biggest players in this space. Prior to OpenAI's launch of Sora, Runway offered some of the most realistic and impressive generative video content, and Gen-3 has a level of motion quality approaching that of Sora, which is very impressive.
Runway was the first to release a commercially available synthetic video model, and over the past year it has added new features and improvements, including highly accurate lip syncing and motion brushes for controlling animation and narration.
With the release of Gen-3, you can now create videos starting from 10 seconds. As it is currently only in Alpha, some of the advanced features like image to video, video to video, clip expansion etc. are not available but are expected to be released soon.
In an increasingly competitive market, Runway remains one of the best AI video platforms, and in addition to generated content, it also offers great collaboration tools and image-based AI features like upscaling and text-to-image conversion.
Runway There is a free plan with 125 credits. The standard plan costs $15/month.
LumaLab Dream Machine
Luma Labs, one of the newest AI video platforms, has come out of nowhere to release Dream Machine, which offers stunning realism, fast tracking, natural motion, and the first five seconds of video generation.
Unlike other platforms, Dream Machine charges one credit per generation, making it easier to track how much you're spending and when you're approaching your limit.
It automatically improves the prompts and ensures a better output. One of the most innovative features is keyframes. You can specify two images (start and end points) and specify how to fill the gap between the two images. This is perfect if you want to create fun transitions or make a character walk across the screen.
The ability to extend clips is also a particularly powerful feature of Dream Machine, allowing you to follow characters or create new scenes. Clips continue from the last frame of the last video, and you can change the motion description for each extension.
Dream Machine There is a free plan with 30 generations per month. The standard plan costs $30 per month and includes 150 generations.
PicaLab
Pika Labs is one of the two big players in the generative AI video space alongside Runway. Their Pika 1.0 model can create videos from images, text, and other videos, and can extend the video up to 12 seconds in length, though the longer it is, the slower it gets.
Pika was released to much acclaim last year with the introduction of a cartoon version of Elon Musk and a nifty inpainting feature that lets you replace or animate specific areas of a clip.
Pika Labs offers negation prompts and fine control over movements within the video, as well as sound effects that are built in from text prompts or that are timed to the video and lip sync.
Pika Labs Lip Sync can be added to video content, for example by generating a video from Midjourney's photos and moving the lips to give it a voice, or, as I did in my experiments, to move an action figure.
We've heard that Pika 2.0 is in the works and has recently introduced significant upgrades to its image-to-video model, improving overall movement and control.
PicaLab There is a free plan with 300 credits. The standard plan costs $10 per month.
Leonardo and the Night Cafe
Because Stable Video Diffusion is an open model, it can be commercially licensed and adopted by other companies, best examples of which are Leonardo and Night Cafe, two AI imaging platforms that offer a range of models that include Stable Diffusion itself.
Branded image platforms Motion by Leonardo and Animate by NightCafe do essentially the same thing: take images you've already created on the platform and animate them. You can set the degree of movement, but other control options are minimal.
Night Cafe The basic plan costs $6 per month for 100 credits.
Leonardo There is a free plan that allows you to create up to 150 images per day, while the basic plan costs $10 per month.
Final Frame
It's something of a dark horse in the AI video space, but it does have some interesting features: A relatively small bootstrap company, FinalFrame is building a “total platform” that competes well with companies like Pika Labs and Runway in terms of quality and features.
FinalFrame gets its name from the fact that it creates the next clip based on the final frame of the previous video, improving consistency across longer video productions. You can generate or import a clip and drop it into the timeline to create subsequent clips or build a complete production.
The startup recently added lip sync and sound effects for select users, including an audio track in the timeline view that lets you add these sounds to your videos.
Final Frame You need to purchase a credit pack that is valid for one month. The basic plan costs $2.99 for 20 credits.
Hyper
A relative newcomer with its own proprietary models, Haiper takes a slightly different approach to other AI video tools, building a foundational model and training dataset that's better suited to following prompts rather than granular control over movements.
In default mode you can't even change the motion levels – you're assuming the AI will figure them out from the prompts, and in most cases it works well – in some testing we found that leaving the motion settings at their defaults worked better than any of the controls you could set.
Hyper It's currently free, and no pricing information has been released.
LTX Studio
Unlike others, it is a complete generative content platform, allowing you to create multi-shot, multi-scene videos from text prompts. LTX Studio has images, videos, narration, music and sound effects, all of which can be generated simultaneously.
The layout is more like a storyboard than the usual prompt boxes and video players on other platforms, and as you generate your video, LTX Studio lets you tweak any element, such as changing the camera angle or bringing in images to animate from an external application.
LTX Studio doesn't handle motion as well as Runway or Stable Video, often resulting in unsightly blurriness and distortion, but these are issues that other tools are starting to solve and are something Lightricks, the owner of LTX Studio, is spending time working on. There is also no lip syncing, but that may be addressed in the future.
LTX Studio There is a waitlist for the beta version, and while no pricing information was provided, the beta version is free to use.