With LTX-2, Lightricks Unleashes AI Video Creation In Full HD

When it comes to creating video content for screens big and small, many companies have a need for speed, which is why they’re increasingly turning to artificial intelligence tools that can accelerate the creative process. In an attempt to meet this demand, Lightricks recently released its fastest-ever AI video model, LTX-2.

The company claims that LTX-2 is capable of creating an entirely original, six-second HD video with just five seconds of compute time. Although slightly slower, the model is also capable of output in Full HD and even 4K resolutions.

LTX-2 also includes improvements when it comes to the accuracy and realism, compared to Lightricks’ LTXV-13B model, while enhancing cost-effectiveness both in terms of its compute efficiency and also through the addition of audio generation. Not only can users spin up high-resolution video, but they can create the corresponding audio track, generating music, dialogue and sound effects at the same time as they bring their visuals to life.

Lightricks says that LTX-2 can be used for “cinematic fidelity,” with support for up to 48 frames per second, making it a viable solution for filmmakers and high-end branded content or virtual effects.

From Consumers To Filmmakers

LTX-2 is blazingly fast, which makes sense given that it comes from a company that has transitioned remarkably quickly itself, switching its attention from the consumer market to encompass professional creators.

The new model is available to access in the LTX Studio platform, which has become the focal point of Lightricks’ shift from the consumer market to business audiences. The company still caters to casual users with its Facetune and Videoleap mobile applications, which are designed for editing and enhancing photos and videos with filters and AI tools, but it has made a real name for itself in the marketing and filmmaking community with its flagship platform.

Lightricks launched LTX Studio early last year, labeling the web app its first professional-grade product. Having made significant advances in AI while experimenting with its consumer apps, the company gained the expertise to cater to professional creatives, such as marketing and advertising teams, aspiring filmmakers and others with ambitions to put out high-end visual content.

LTX Studio is an AI filmmaking platform that aims to reduce friction from every aspect of the content creation process. Powered by the company’s open-source LTXV models but integrating third-party AI tech such as Nano Banana, Flux and Veo, LTX Studio gives users access to capabilities such as storyboarding, camera controls, character and actor management, motion capture and the ability to upload full scripts and reference images, so they can craft production-ready video content.

The platform has been drawing significant attention, both from businesses and the media. Earlier this month, seasoned marketing professionals from three agencies: Code and Theory, Instrument and Kettle, participated in Lightricks’ VibeSprint, a high-stakes experiment in AI-driven creativity. The teams were tasked with using LTX Studio to transform a brief into an entirely fresh video marketing campaign in just two days.

The event showcased how AI can help professional teams to accelerate their work, freeing them from the more time-consuming aspects of content creation, so they can focus on experimentation and shaping their ideas.

Still, Lightricks hasn’t forgotten its consumer roots. Over the last year, the company has fused many of the capabilities of its AI models into its Facetune and Videoleap apps, growing their popularity. According to published figures, the company’s applications have been downloaded more than 730 times, while their user base has grown to over 6.6 million subscribers.

Open-Source Access

With the launch of LTX-2, it’s clear that Lightricks is aiming to serve as many audiences as it can. While LTX Studio is for professionals, the model itself is open-source and free to download, so everyone has the ability to generate top-notch video content starting with nothing more than a prompt.

In order to make LTX-2 accessible, Lightricks has also focused on making it as efficient as possible. And while it may lose some of its processing speed, it’s perfectly capable of generating the same, high quality video on a consumer grade GPU as it is on a cluster of more advanced chips. Economics have long been a focus for the company, claiming that LTX-2 can run at up to 50% lower costs than rival models such as OpenAI’s Sora.

Customers who access the model via LTX Studio or Lightricks’ API will have to pay for usage, with prices starting at just four cents per second for the fastest iteration, though they’ll be limited to HD video only. There’s also a balanced pricing tier at seven cents per second, while the Ultra version generates video more slowly, but ups the resolution to 4K and also creates the sound to go with it.

“This isn’t vaporware or a research demo,” said Lightricks Chief Executive Zeev Farbman in a statement. “It’s a real breakthrough in video generation. It’ll be made open-source this fall and built for production, from developers building new tools to studios and enterprises delivering final 4K content.”