LTX 2.3 is a 22-billion-parameter open-source AI video generator that converts text, images, or audio into cinematic video clips using a Diffusion Transformer architecture built by Lightricks.
What is LTX 2.3?
LTX 2.3 is an open-source AI video generation model with 22 billion parameters, developed by Lightricks. It takes text prompts (up to 2,000 characters), images, or audio as input and outputs high-definition video clips with fluid motion, accurate lighting, and natural physics. The model runs on cloud servers via ltx23.app — no local GPU required — and is also available for local inference with ComfyUI workflows and quantized formats on Hugging Face.
Key Features
- 22 Billion Parameter DiT Engine — Built on Diffusion Transformer architecture delivering sharper textures, finer edges, and better detail than prior versions.
- Expanded Text Connector (4x) — A 4x-larger text connector interprets complex prompts with spatial layout, character actions, and mood accurately in one pass.
- Native Portrait Video at 1080×1920 — Create vertical video natively trained on real portrait data, ideal for Reels, Shorts, and TikTok.
- Multi-Modal Generation — Supports text-to-video, image-to-video, audio-to-video, video-to-video, and depth conditioning from one unified platform.
- Rebuilt VAE for Sharper Output — Reworked latent space produces crisper hair, cleaner edges, and better texture preservation at every resolution.
- Face & Character Preservation — Consistent faces, expressions, and body proportions across frames, ideal for storytelling and multi-shot sequences.
- 18x Faster than WAN 2.2 — On H100 GPUs, LTX 2 series achieves roughly 18x the throughput of WAN 2.2 14B, enabling rapid batch rendering.
- Open Source & Free to Use — Weights are open-source on Hugging Face; free for personal and commercial use under $10M annual revenue. On ltx23.app, new accounts receive free credits to try.
Who is it for?
- Social media managers — generate vertical AI video clips for TikTok, Instagram Reels, and YouTube Shorts natively at 1080×1920.
- Filmmakers and content creators — produce cinematic b-roll, storyboard animations, and product demos without traditional editing workflows.
- E-commerce teams — create product videos at catalog scale using image-to-video generation, reducing studio costs.
- Marketing and advertising teams — localize ad campaigns for multiple markets with audio-to-video synchronization and consistent branding.
What can you do with LTX 2.3?
- Text-to-video: Describe any scene and LTX 2.3 generates a cinematic clip with fluid motion, accurate lighting, and natural physics in seconds.
- Image-to-video: Upload a reference image and LTX 2.3 transforms it into a dynamic video clip with smooth camera motion and lifelike animation.
- Audio-to-video: Feed an audio track and LTX 2.3 generates matching video with tight lip-sync, beat-aligned motion, and spatial audio cues.
- Video-to-video and depth conditioning: Apply stylistic changes or depth-based effects to existing video clips using the unified pipeline.
How does LTX 2.3 work?
- Enter your prompt — describe the video in natural language (up to 2,000 characters) or upload an image or audio clip as reference.
- Set parameters — choose video duration (4–20 seconds), aspect ratio (16:9, 9:16, 1:1, 4:3), and output quality up to 1080p HD.
- Generate & download — click generate; the cloud renders the video moments later, ready to download with full commercial rights.
Pricing
LTX 2.3 offers a freemium model with free credits for new accounts. Paid subscriptions include:
- Starter: $19.9/month ($13.9/month billed yearly) — 14,400 credits/year, includes all models, up to 4K resolution, commercial use license.
- Premium: $39.9/month ($27.9/month billed yearly) — 33,600 credits/year, priority customer support.
- Advanced: $99.9/month ($69.9/month billed yearly) — 120,000 credits/year, fastest generation speed, expert team support. Cost per 100 credits ranges from $0.70 (Advanced) to $1.16 (Starter). A 30% discount is available for a limited time.
FAQ
Is LTX 2.3 free to use?
Yes. New accounts on ltx23.app receive free credits to try AI video generation. After that, you can purchase additional credits or subscribe to a plan. The open-source model weights are free for commercial use under $10M annual revenue.
What hardware do I need to run LTX 2.3 locally?
For local inference, recommended setup is an NVIDIA GPU with 32 GB+ VRAM, 32 GB RAM, and 60 GB storage on Windows. Alternatively, use ltx23.app for cloud rendering with no local GPU required.
Can I use LTX 2.3 outputs commercially?
Yes. Videos generated on ltx23.app include full commercial rights — no watermarks, no royalty fees. The open-source license also permits commercial use of locally generated outputs for qualifying organizations.
What video formats and resolutions are supported?
Videos render at up to 1080p HD in aspect ratios 16:9, 9:16 (native portrait 1080×1920), 1:1, and 4:3. Duration ranges from 4 to 20 seconds including audio-synced output.
How does LTX 2.3 compare to other models like WAN 2.2?
On H100 GPUs, LTX 2 series achieves approximately 18x the throughput of WAN 2.2 14B. LTX 2.3 also introduces native portrait video, a rebuilt audio vocoder, and sharper edge detail from its reworked latent space.
