🇮🇱 Israel🎬 Video generationReleased March 2026
LTX-2.3
Lightricks
Open-sourceSelf-hostRGPD
Verdict
A solid pick for self-hosting and full data control.
Overview
LTX-2.3 is a DiT-based diffusion model that generates synchronized video and audio in a single pass. Its distilled 8-step checkpoints and LoRA variants let teams fine-tune motion, style or voice in under an hour. Prompt adherence remains inconsistent, and audio without speech tends to sound thin. Open weights under Apache 2.0, runs entirely on a desktop-class machine with no cloud dependency.
Skill profile
Not disclosed
Strengths
- Open-source and self-hostable
Limitations
- API pricing not disclosed
Who is it for
A good fit if…
- you want to control cost or self-host
- you have GDPR constraints
Skip it if…
Ideal use cases
- Synchronized video prototyping
- Fast marketing content
- Fine-tuning style or voice
- Local generation without cloud
- Custom ComfyUI pipelines
Access & availability
Self-hostable (open weights)
Key specifications
Context
0
Input price
—
Output price
—
Speed
—
Price not auditedBenchmarks not audited
Privacy
GDPR-compliantHosting : self
Run it locally
Desktop with solid GPU / 32GB+ Mac1.2M2k
| Quantization | Disk | RAM / VRAM | Typical hardware |
|---|---|---|---|
| Q4 · recommended | 12.5 GB | 16 GB | 32GB PC / 24GB Mac / RTX 4070 Ti+ |
| Q8 · balanced | 23.5 GB | 28 GB | 64GB Mac / dual 24GB GPUs |
| FP16 · max quality | 44 GB | 50 GB | 64GB Mac / dual 24GB GPUs |
Estimates for a moderate context. Long contexts need more RAM (KV cache).
Deploy
Copy-ready commands generated from this card. Adjust context length and GPU count to your hardware.
Estimated memory · FP16 50 Go · Q4 16 Go
OpenAI-compatible server for production on NVIDIA GPUs.
pip install vllm
vllm serve Lightricks/LTX-2.3 \
--max-model-len 32768 \
--tensor-parallel-size 1 \
--dtype autoLightricks/LTX-2.3
Advanced data · for expertsArchitecture, modalities, detailed cost, full benchmarks▾
Advanced data · for experts
Architecture, modalities, detailed cost, full benchmarks
Architecture
dense
Size
22B
Cutoff date
—
Inputs
text
Outputs
text
License
apache-2.0
Hosting
self
Hallucination score
—
Estimated cost (API)
Typical exchange (~3k in / 1k out)
Not disclosed
1M in + 1M out
Not disclosed
All benchmarks
| Arena Elo | — |
| MMLU | — |
| GPQA | — |
| HumanEval | — |
| SWE-Bench | — |
| MATH | — |