LTX-2.3 22B
Lightricks/LTX-2.3
published Mar 2026 · updated Apr 2026
LTX-2.3 is an open-source DiT-based audio-video foundation model that generates synchronized video and audio from text prompts.
specs
| Task | Audio-Video Generation |
| Architecture | Diffusion Transformer (DiT) |
| Parameters | 22B |
| Language | English |
best for
- ·Generating synchronized video and audio from text prompts for short films
- ·Creating background music and sound effects for video clips
- ·Rapid prototyping of audiovisual content
FAQ
It accepts text prompts in English and generates synchronized video and audio as output.
LTX-2.3 has 22 billion parameters (22B).
The model is released under an open-source license; refer to the LICENSE file on Hugging Face for exact terms.
Use the gigarouter OpenAI-compatible endpoint with your API key to run inference.
The base model primarily generates video from text. Image-conditioned versions are available via LoRA adapters (IC-LoRA).
We're benchmarking and onboarding LTX-2.3 22B as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.