LTX 2.3 Pro example output

LTX 2.3 Pro

Open source video with native audio. Premium 6–10 second clips up to 4K.

LTX 2.3 Pro from Lightricks is a 22-billion parameter open-source video model generating up to 4K (2160p) video with synchronised audio in a single pass. Ambient sound, dialogue, and effects are produced alongside the video rather than layered afterward. Duration locks to 6, 8, or 10 seconds at the standard endpoint, with three resolution tiers — 1080p, 2K, and 4K. The rebuilt VAE delivers sharper fabric, cleaner hair, and stable reflections during camera moves.

Example outputs

LTX 2.3 Pro example 1

A slow dolly around a contemporary museum building, angular white concrete facades catching late afternoon sun, the reflecting pool mirroring the structure, architectural film

LTX 2.3 Pro example 2

Camera pushing in on a luxury perfume bottle as mist swirls around the base, golden liquid catching dramatic studio light, premium product commercial

LTX 2.3 Pro example 3

A cinematic flyover of floating fantasy islands connected by ancient stone bridges, waterfalls cascading into clouds below, a lone traveller crossing, epic game trailer

LTX 2.3 Pro example 4

A smooth walkthrough of a boutique hotel lobby, camera gliding past the marble reception desk and under hanging pendant lights, warm evening atmosphere, interior showcase

LTX 2.3 Pro example 5

Camera circling a custom cafe racer motorcycle on rain-wet cobblestones at night, chrome and leather catching streetlamp glow, moody automotive reveal

LTX 2.3 Pro example 6

A fashion model mid-stride on a runway, camera tracking alongside, dramatic spotlights cutting through haze, front row silhouettes, fashion show footage

How it works

01

Describe your scene

Type a detailed prompt describing the video you want, or upload a reference image as a starting frame.

02

Choose your settings

Pick your resolution and duration. See the credit cost before you generate.

03

Generate your video

Your video is ready in 1-3 minutes. Download, iterate, or extend the sequence.

Ready to create with LTX 2.3 Pro?

Jump into the Studio and start generating. Plans from £10/month.

Open-source 4K video with native audio in one pass.

LTX 2.3 Pro from Lightricks is one of the most capable open-source video models available. At 22 billion parameters, it generates up to 4K (2160p) video with synchronised audio in a single forward pass. Pick between 6, 8, or 10 seconds at any of three resolution tiers — 1080p, 2K (1440p), or 4K (2160p).

The rebuilt VAE with a redesigned latent space means sharper fabric, cleaner hair, and stable chrome reflections during camera moves. A 4× larger text connector allows complex multi-subject prompts to hold across the full clip.

LTX 2.3 is the first production model to generate native 9:16 portrait video without cropping, making it ideal for short-form social content on TikTok, Instagram Reels, and YouTube Shorts. For longer clips up to 20 seconds at the same resolution range, switch to LTX 2.3 Fast — same underlying model, extended duration ceiling.

Native portrait mode

Generate vertical 1080×1920 video natively. No cropping from landscape. The model understands vertical composition and places subjects correctly for mobile-first content.

Audio in one pass

Sound effects, ambient audio, and dialogue are generated alongside the video. Timing is determined by the model, not post-production alignment.

Premium short-form ceiling

Pro is locked to 6, 8, or 10 second outputs at the standard endpoint. For 12–20 second clips at the same resolution range, the Fast variant reaches longer durations using the same 22B-param architecture.

Three resolution tiers

1080p at $0.06/sec, 2K (1440p) at $0.12/sec, 4K (2160p) at $0.24/sec. Clean 2× scaling per tier so the cost decision tracks the resolution decision directly.

Frequently asked

Questions about LTX 2.3 Pro.

LTX 2.3 Pro from Lightricks is a 22-billion parameter open-source video model generating up to 4K (2160p) video with synchronised audio in a single pass. Ambient sound, dialogue, and effects are produced alongside the video rather than layered afterward. Duration locks to 6, 8, or 10 seconds at the standard endpoint, with three resolution tiers — 1080p, 2K, and 4K. The rebuilt VAE delivers sharper fabric, cleaner hair, and stable reflections during camera moves.
Built differently

Why Stensyl?.