LTX-2.3

by Lightricks

Video + audio generation model by Lightricks. 21B parameters, Diffusion Transformer architecture. Text-to-video and image-to-video, synchronized audio, up to 4K. LTX-2 Community License.

VideoLTX-2 Community LicenseReleased Mar 202621B parameters

Hugging Face repo ↗

Drag or click to move along the pipeline

full qualitylightest

Lighter tiers fit smaller GPUs and lose some detail. Start with the heaviest one that loads.

The model

LTX-2.3

BF1669.5 GB

Three pieces, three separate downloads.

Diffusion modelGoes in checkpoints/

Generates the image

43.0 GB↓↗

Text encoderGoes in text_encoders/

Turns your prompt into guidance

Pick one: smaller loads faster, larger keeps more detail.

22.7 GB↓↗

2.1 GB↓↗

VAEGoes in vae/

Decodes the result into pixels

Pick one: smaller loads faster, larger keeps more detail.

1.4 GB↓↗

0.3 GB↓↗

LoRAs (optional)Goes in loras/

7.1 GB↓↗

SamplerKSampler

Copy these values into your KSampler.

CFG1.0 (distilled) · 3.0-3.5 (full)

Frame_rate24 fps

Max_duration257 frames (~10.7s)

Lora_strength0.5

Steps_stage_18 (distilled) · 30-40 (full)

Steps_stage_23 (distilled) · 10-15 (full)

Vae_tile_size768

Default_duration121 frames (~5s)

Default_resolution1088×1344 (portrait)

Usage notes

GGUF is the lightest way to run this model.

LTX-2 Community License. Check the official repo for usage terms.

Click a file name to copy it.