Lightricks, the Israeli tech company best known for apps like Facetune and Videoleap, is expanding into professional production of AI generative videos that will set them apart from the competitors
With the launch of its new model, LTXV, the company claims it can produce continuous AI-generated video content exceeding 60 seconds. This is significantly longer than what’s currently possible with leading models like OpenAI’s Sora, Google’s Veo or Runway’s Gen-4..
According to Lightricks co-founder and CEO Zeev Farbman, LTXV “unlocks a new era for generative media.” He explains that the model is designed to start streaming results immediately, generating the first second almost instantly and building the rest of the sequence on the fly.
Story continues below this ad
The system uses overlapping frame chunks to preserve continuity—ensuring that characters, motion, and storyline remain consistent across time. This autoregressive approach is similar to the way large language models like ChatGPT generate text, except LTXV applies it visually, frame by frame.
LTXV reportedly delivers output much faster than competitors like Veo 3, Runway Gen-4, or Kuaishou’s Kling, which often make users wait several minutes for a few seconds of video. In a live demo, Lightricks showcased a continuous 60-second video featuring a woman cooking as a gorilla walks in and hugs her, an example of how the model maintains narrative flow without stuttering or abrupt transitions.
That LTXV is open source and unrestricted by a proprietary API is notable. The model will be available on GitHub and Hugging Face as open weights. It is free to use for individuals and small teams earning less $10 million annually.
According to Farbman, this supports Lightricks’ “open development for real-world application” approach, which gives developers and independent artists the freedom to expand upon the core engine.
Story continues below this ad
From a technical standpoint, the new model is fast and lightweight. It can be powered by a single Nvidia H100 or even high-end consumer GPUs. However, Farbman points out that available benchmarks for other models often require multiple H100s to produce just five seconds of high-resolution video.
Veo 3 from Google LLC is the only AI video model that can also create its own audio tracks. Nevertheless, Lightricks’ most recent developments coincide with the big AI video production companies. Everyone is making efforts to set themselves apart from the competition, and its rivals can boast a number of distinctive features of their own.
© IE Online Media Services Pvt Ltd
Expand
