LTX-2: Two New Models from Lightricks

The world of AI video generation is moving at an incredible pace. Just when the creative community gets its hands on one groundbreaking model, another emerges, pushing the boundaries of what's possible. In this rapidly evolving landscape, Lightricks, a company with a long history of building powerful creative tools, has introduced LTX-2, a next-generation open-source foundation model designed not just to generate stunning visuals, but to integrate seamlessly into real-world production workflows.

LTX-2 isn't just another text-to-video generator; it's positioned as a complete AI creative engine. With two distinct models, LTX-2 Fast and LTX-2 Pro, it aims to serve the entire creative spectrum—from rapid ideation on a mobile device to producing high-fidelity, polished footage for professional projects. This article dives deep into what makes LTX-2 a significant new player, exploring its technology, capabilities, and who stands to benefit most from this powerful new tool.

Table of Contents

• A New Creative Engine from a Veteran Company
• Technology Overview: Built for Production
• A Look Under the Hood: Model Stats
• Core Capabilities: Strengths and Weaknesses
• Showcase: LTX-2 in Action
• Who Is LTX-2 For?
• Pricing, Performance, and Accessibility
• Frequently Asked Questions
• Related Reading
• Sources & Transparency
• Article Metadata

A New Creative Engine from a Veteran Company

LTX-2 is developed by Lightricks, a company that has been at the forefront of the creator economy for over a decade. Founded in 2013, Lightricks is the force behind globally popular photo and video editing applications like RunDiffusion, Facetune, Videoleap, and Photoleap. With a mission to build innovative tools that bridge the gap between imagination and creation, the company has empowered hundreds of millions of users to create and share compelling visual content.

This deep experience in building user-centric creative software provides crucial context for LTX-2. Lightricks understands the pain points of creators and production studios. They know that while jaw-dropping quality is important, it's useless without speed, control, and the ability to integrate into an existing pipeline. LTX-2 is the culmination of this experience—an AI model built not in a research vacuum, but with the practical needs of artists, developers, and studios in mind.

Technology Overview: Built for Production

At its core, LTX-2 is an open-source foundation model designed for generating video and synchronized audio. Lightricks describes it as the "most complete AI creative engine in the world," a bold claim backed by a focus on production-ready features. The architecture is built on principles of speed, efficiency, fidelity, and openness.

One of its most ambitious promises is the ability to deliver native 4K video at 50 frames per second (fps) for durations of up to 20 seconds. This is a significant leap forward, as many existing models struggle to generate high-resolution content without requiring extensive upscaling and post-processing.

The LTX-2 ecosystem is designed around an API that exposes three distinct "flows" tailored to different needs:

1. LTX-2 Fast: Optimized for extreme responsiveness and rapid iteration. This model is ideal for storyboarding, previz, and mobile workflows where speed is more critical than perfect visual fidelity.
2. LTX-2 Pro: The workhorse model that balances high-quality output with efficient performance. This flow is designed to produce polished, professional-grade video suitable for marketing, social media, and independent film projects.
3. LTX-2 Ultra: A forthcoming model that will prioritize maximum fidelity for the most demanding use cases, such as VFX and cinematic productions.

A key strategic advantage is its open-source nature. By providing transparent access to model weights, Lightricks is empowering developers to build custom workflows, fine-tune the model on specific data, and contribute to a community-driven ecosystem. This approach also allows for a unique Local → Cloud workflow, where creators can iterate quickly on consumer-grade GPUs before pushing final renders to more powerful cloud hardware.

A Look Under the Hood: Model Stats

Lightricks has focused its initial announcements on the practical capabilities and workflow benefits of LTX-2 rather than releasing a detailed technical paper. As such, specific architectural details like the number of parameters or the exact composition of the training dataset have not been made public.

Note: This information is based on the public API preview and official announcements. As of October 2025, more detailed specifications are expected to be released.

Core Capabilities: Strengths and Weaknesses

No model is perfect for every task. LTX-2 makes clear trade-offs to excel in its target areas. Here’s a breakdown of its key strengths and current limitations.

Strengths

• Production-Grade Output: The headline feature is the native 4K at 50 fps for 20-second clips, first in the world and LTX is leading the charge.. This moves AI video from a novelty to a viable tool for professional use. Finally.
• Efficiency and Cost: Lightricks claims LTX-2 can be up to 50% more cost-effective than competing models. Its ability to run on consumer-grade GPUs lowers the barrier to entry for smaller creators and studios.
• Open and Flexible: As an open-source model with a robust API, LTX-2 is built for integration. This allows developers to incorporate it into existing software (like editing suites or VFX pipelines) and build entirely new applications on top of it.
• Creative-Centric Design: The Fast and Pro model distinction shows a deep understanding of the creative process. It acknowledges that iteration and final output have different needs, providing a tool for each stage.
• Integrated Audio: The promise of synchronized, native audio generation within the same model is a game-changer, potentially eliminating the difficult and time-consuming step of scoring and sound-designing AI-generated clips.

Weaknesses

• Audio: Native audio generation is something to be desired. We anticipate this could be getting better with future releases.
• Growing Ecosystem: As a new entrant, the community, tooling, and number of integrations are still developing. While being open-source is a massive advantage, it will take time to build the rich ecosystem that more established platforms enjoy.
• Competition: LTX-2 enters a fiercely competitive market with major players like OpenAI's Sora, Google's Veo 3.1, and Kling 2.5 Fast and Kling 2.5 Standard. It will need to deliver on its promises of efficiency and workflow integration to carve out a significant market share.

Showcase: LTX-2 in Action

The true test of any generative model is the quality and diversity of its output. The following examples, generated with the LTX-2 Pro model, showcase its ability to handle a wide range of styles, from photorealistic scenes to imaginative animations.

Prompt: "In a tender cinematic moment, a young boy stands quietly by a rain-speckled window, his face lit by the glow of fireworks bursting in the evening sky. The camera lingers on his thoughtful profile as colorful explosions of red, green, and gold reflect faintly across the glass. Outside, the celebration feels distant, while inside, his stillness conveys both wonder and melancholy, capturing the quiet contrast between noise and solitude."

Prompt: documentary realistic, In a dimly lit bedroom, a woman in lilac pajamas sits by the window, shadows of blinds cutting across her face. Suddenly, beams of car headlights sweep past, casting shifting light through the room. She squints, unsettled, then slowly reaches for her phone on the bed. The glow of the screen illuminates her features as she stares at it with tense anticipation, the fleeting headlights fading into silence."

Prompt: In a dimly lit café on a rainy evening, a young woman sits near the window, her expression tense as raindrops trace delicate lines down the glass. The blurred glow of streetlights outside reflects faintly on her face, highlighting the unease in her eyes. Suddenly, a man in a butter-yellow bomber jacket enters the frame and takes the seat across from her. We only see him from behind — the padded shoulders of the jacket, the sturdy collar, the weight of his presence filling the space. As he sits, she stiffens, her lips part slightly in silent tension, her gaze avoiding his while betraying her discomfort. The camera holds on her strained expression, contrasting the muted hum of the café with the heavy silence between them, creating an atmosphere thick with suspense and unanswered questions.

Prompt: A medium shot captures a young woman standing in a field of tall, golden grass, holding a rifle pointed forward. She wears a red and blue plaid shirt over a grey t-shirt. The sun is setting, casting a warm glow over the scene. Behind her, out of focus, is a dilapidated wooden house. The camera begins to orbit behind her, moving from her right to her left, keeping her centered in the frame. As the camera moves, the view shifts to show her from behind, her brown hair slightly blowing in the breeze. The camera then shifts in focus fro her to a slender dark figure standing in front of her in the distance, seen as a silhouette.

Prompt: A high-speed drone shot captures a surreal scene in a dimly lit subway station, where a woman is flying effortlessly above the ground. She is wearing a flowing, traditional garment with red floral patterns, which billows dramatically around her. Her long, dark hair streams back in the wind, mirroring the rapid forward motion. The camera tracks her from the front, aligned perfectly to her unwavering, determined gaze. The background streaks past in a blur, with tiled walls and structural pillars creating a tunnel-like effect. Fluorescent lights overhead provide a stark, linear illumination, casting moving shadows across her face and highlighting the movement. The platform edges, outlined by fluorescent yellow safety lines, zip by beneath her, while distant, indistinct passengers and signs add to the vivid sense of speed and direction. The shot maintains the woman’s steady, surreal flight through the bustling transit space, creating an atmosphere of dynamic elegance and otherworldly grace.

Prompt: The camera opens on a young woman standing in front of an abandoned gas station, her face frozen in shock. She suddenly raises her trembling hands to cover her mouth, eyes widening in sheer horror. The sound of crackling flames grows louder as the camera holds on her terrified expression for a moment. Then, in one fluid motion, the camera rotates around her shoulder to reveal what she is staring at: a car engulfed in fire, thick smoke billowing into the cold air. The flames roar violently, casting an orange glow on the desolate surroundings, contrasting with her pale, frightened face. The shift from her silent panic to the shocking sight of the burning vehicle heightens the sense of dread and cinematic drama, immersing the viewer in the unfolding chaos.

Prompt: documentary realistic shot, At sunset, two young lovers sit on the hood of a car in a quiet suburban street, the sky painted in soft pink and violet tones. The camera lingers in a gentle close shot as the boy leans in slowly toward the girl, their faces silhouetted against the fading light. His movements are tentative, full of anticipation, as he lowers his head closer, ready to kiss her while she tilts her chin up expectantly

Prompt: A medium shot focuses on the upper body of an old lighthouse guard with grey hair and a beard, wearing a dark jacket and a black turtleneck sweater. He stands on a grassy area in front of a white lighthouse that features a red top. The scene is set on a rainy day, with raindrops visibly falling. The man turns around and walks away slowly, showing him from behind standing on the edge of a cliff in front of the stormy ocean.

Prompt: A man in a burgundy suit leans over an open car trunk at sunset, his intense eyes locked on the viewer. Slowly, he reaches inside with one hand and pulls out a pistol, the metallic surface glinting faintly in the warm light. His other hand presses firmly on the edge of the trunk as he straightens his posture, exuding calm menace. After a tense pause, he lowers his gaze, then slams the trunk shut with deliberate force, the camera catching the sharp motion and the final echo of the closing lid. The atmosphere is cinematic, suspenseful, with the glowing diner sign in the background contrasting the darkness of the act.

Prompt: Two young men stand close, gripping each other with intense emotion. Slowly their struggle softens, their hands shifting from the other’s face and neck into an embrace. Their heads lower against each other’s shoulders as tears begin to fall. They shake slightly, crying quietly, holding tightly as if not wanting to let go. The sunlight flickers across the corrugated metal wall behind them, grass swaying in the breeze, as the camera holds steady on the raw, emotional moment.

Prompt: The little girl crouches on the grass, pressing two ripe tomatoes to her cheeks with a playful puffed expression. She smiles and blinks softly, tilting her head a touch. A gentle breeze lifts her blue headscarf and ruffles the checkered apron. Sunlight flickers on the clear water bucket, tiny ripples shimmering. Camera holds a steady close-up, warm, natural daylight.

Prompt: A medium shot in a library, where a man and a woman are browsing through books on opposite sides of a shelf. The man, on the left, is wearing a light beige sweater and is looking at the records. The woman, on the right, is wearing a white long-sleeved shirt and a black skirt, holding a book in her hands. She turns around, a smile forming on her face, and shows the man something funny on the book she's holding. The man chuckles, his eyes crinkling at the corners as he looks at the record. They both begin to walk slowly towards the camera.

Prompt: A cinematic close shot of a rhythmic gymnast by the balance beam, bathed in shafts of light from tall gym windows. The camera begins slightly above her shoulder, then glides slowly in a gentle arc around to her front, revealing her focused expression. As the camera moves, she performs a subtle motion: stretching her arms outward, then lifting her chin softly toward the light. Dust particles shimmer in the air, adding texture and atmosphere. The mood is quiet, powerful, and poetic, highlighting strength and grace with restrained movement.

Prompt: A medium shot shows a man with short grey hair, wearing a dark blue coat over a grey sweater, walking on a bridge. The camera tracks his walk, keeping him centered in the frame. The background features a body of water with docked sailboats and buildings under an overcast sky. suddenly the camera crash zooms on his face, the man's expression shifts to alarm, his eyes widening and mouth opening slightly as he glances to his left. He abruptly turns and begins to run in the opposite direction. The camera smoothly tracks his run, maintaining a consistent distance. His coat flaps behind him as he picks up speed, his arms pumping and head turning slightly from side to side as he flees, his footsteps echoing on the wooden planks of the bridge. The background blurs slightly due to the camera's movement, emphasizing the man's panicked escape. He continues running until the end of the shot, his figure receding into the distance.

Prompt: "In a realistic cinematic style, a woman stands at the doorway of her house, holding a letter with tense hands. She reads in silence, her brows furrowing as the sunlight falls across her face. After a long pause, she lifts her gaze, staring forward in thought. With a heavy breath, she folds the paper slightly, turns, and steps back inside the house, the door closing softly behind her, leaving the yard empty in fading light."

Prompt: "On the dimly lit porch of a worn wooden house, a man sits beside his dog as the warm glow of a lamp spills from the window. The quiet is broken when the dog suddenly stirs, ears twitching before it rises and begins to bark into the night. The man reaches out calmly, stroking its head and back, whispering softly as his hand steadies the animal, easing its tension while the silence of the evening slowly returns."

Who Is LTX-2 For?

LTX-2 is designed to be a foundational building block for a wide array of users, from individual creators to large-scale enterprises.

• Independent Creators & Developers: The open-source model and consumer hardware compatibility make it accessible. Developers can build custom tools, while creators can generate high-quality footage without needing a render farm.
• Marketing and Advertising Teams: The LTX-2 Pro model is perfect for creating polished social media content, product videos, and ad campaigns quickly and affordably. The ability to iterate rapidly with the Fast model is a major workflow enhancement.
• VFX and Animation Studios: For larger studios, LTX-2's API-first approach allows it to be integrated as a component in a larger pipeline. It can be used for previz, generating background plates, or even creating final animated sequences.
• Enterprise and Production Houses: The focus on efficiency, scalability, and production workflows makes LTX-2 a compelling option for major studios looking to leverage AI to reduce costs and accelerate timelines.

Pricing, Performance, and Accessibility

During the initial preview phase, access to the LTX-2 API is available through a waitlist. Lightricks has not yet published a public pricing page, which is common for new models rolling out access gradually.

However, a core message in their positioning is affordability and efficiency. The company claims that LTX-2 can be up to 50% more cost-effective than other leading AI video generation models. This, combined with its performance on consumer-grade hardware, suggests a pricing model designed for broad adoption.

Interested users can sign up for the API preview on the official Lightricks website to get access and receive details on pricing as they become available.

Frequently Asked Questions

Sources & Transparency

This article is based on publicly available information from Lightricks, including official announcements and API documentation. As of October 2025, a detailed technical paper with model architecture and training data specifics has not been released.

• Official Website: Lightricks
• LTX-2: Introducing LTX-2 by Lightricks
• LTX-2 API Docs: LTX-2 API Docs