LTX-2 Surpasses Sora 2, Defining the 20-Second AI Video Era

Introduction

While we were still in awe of Sora 2 extending AI video to 12 seconds, LTX-2 has once again shattered this boundary — directly pushing video generation to 20 seconds.

This is not just an extension of time, but a leap in creative possibilities. And now, this powerful tool is within reach.

Today, we are extremely excited to announce that LTX-2, the next-generation AI video model from Lightricks, has now landed on WaveSpeedAI.

It is said that LTX-2 introduces three distinct editions — Fast, Pro, and Ultra — to meet every creator’s needs across speed, cost, and visual fidelity.

20-Second Continuous Generation

Say goodbye to short, looping clips. While Sora 2 caps at 12 seconds, LTX-2 supports up to 20 seconds of seamless, story-driven video — allowing creators to build complete sequences from a single prompt.

This additional duration makes all the difference. Whether it’s a product demonstration, a movie trailer, or a character close-up, LTX-2 can maintain consistency in motion, expression, and lighting from start to finish.

This is not just longer video — it’s longer storytelling.

Real Usability

Compared with Sora 2, LTX-2 delivers a full cinematic production line right to your screen.

In side-by-side tests — such as the skydiver scene captured from a front-facing action cam — LTX-2 produces characters that feel alive: faces react naturally, lips sync perfectly with speech, and the surrounding soundscape moves with the wind and altitude.

In contrast, the characters in Sora 2 appear emotionally stiff, lacking the authenticity of the real world.

• Generated by Sora 2:

• Generated by LTX-2:

Professional Creative Control

LTX-2 understands camera language, lighting, and texture — letting you direct your shots instead of rolling the dice. Your prompt becomes more like a shot instruction than a wish list.

Example prompt:

Tracking shot behind a runner as sunlight spills over the city skyline. Soft morning light, lens flare, and natural motion blur.

LTX-2 interprets every cue — the direction of light, the rhythm of movement, even how the camera breathes with the subject — to produce a shot that feels cinematic and intentional.

Flawless Style Adaptation

LTX-2 intelligently adapts to any artistic style while maintaining flawless subject coherence.

Whether your reference is photorealistic or stylized, the model translates it seamlessly into expressive motion.

Prompt Example:

A beautiful anime-style Japanese high school girl sitting under cherry blossom trees. Soft petals fall gently in the breeze. She wears a sailor school uniform with a gentle and slightly shy expression. Warm sunlight filters through the blossoms, casting soft highlights on her face. The background is blurred with cinematic depth of field. The camera slowly zooms in toward her face, capturing the delicate play of light and motion. The overall tone is romantic, calm, and pastel — with a Japanese spring aesthetic and natural lighting.

Conclusion

LTX-2 marks a new era in AI video creation — transforming it from a demo tool into a true production platform.

With 20-second cinematic storytelling, precise camera control, and seamless style adaptation, it empowers creators to move from idea to execution in seconds.

Fast, affordable, and remarkably real — now, the only limit is your imagination.

Try LTX-2 on WaveSpeedAI Today

Experience the next generation of AI video:

• LTX-2 Fast (Text to Video)
• LTX-2 Pro (Text to Video)
• LTX-2 Fast (Image to Video)
• LTX-2 Pro (Image to Video)

Start creating your first AI video masterpiece now on WaveSpeedAI — no waitlist, no code, just creation.

Stay Connected

Join Discord Community | Follow us on X (Twitter) | Open Source Projects

© 2025 WaveSpeedAI. All rights reserved.