Local video-generation pipelines
Run, modify, and extend an audio-video model on owned infrastructure with available weights and workflows.
Independent tool overview
LTX-2.3 at a glance
LTX-2.3 is Lightricks' open-weight multimodal video model for generating and editing synchronized video and audio locally or through an API, with native portrait, up to 4K API output, HDR, and production controls.
Visit the official LTX-2.3 site ↗
- Developer
- Lightricks
- Model type
- Open-weight audio-video diffusion transformer
- Variants
- LTX-2.3 Fast and Pro
- API output
- Portrait or landscape, 720p to 4K
- Newer family
- LTX-2.5
Overview
What LTX-2.3 is
LTX-2.3 upgraded the LTX-2 family with sharper detail, stronger prompt adherence, more motion in image-to-video, cleaner synchronized audio, and native vertical generation. It supports text-to-video and image-to-video through Fast and Pro variants, while Pro also handles audio-to-video, retakes, extensions, reframing, and HDR workflows.
The model remains available for self-hosting and through the managed LTX API, even though LTX-2.5 is now the newer family. That makes this page most relevant to teams evaluating the proven 2.3 checkpoints, maintaining an existing integration, or using 2.3-only editing endpoints and local workflows.
Use cases
Who LTX-2.3 is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
Developer API products
Integrate text, image, or audio-conditioned video without hosting the model stack.
Vertical social video
Generate native 9:16 footage trained for portrait orientation instead of cropping landscape output.
Controlled video editing
Retake a selected section, extend a clip, reframe footage, or convert SDR footage to HDR.
Capabilities
Core LTX-2.3 features
Joint audio-video generation
Produces synchronized visuals and sound in one multimodal model instead of adding audio as a separate afterthought.
Text, image, and audio conditioning
Supports text-to-video, image-to-video, and Pro-only audio-to-video workflows.
Native portrait and higher resolutions
The API supports portrait and landscape generation from 720p through 4K, depending on model and endpoint.
First-to-last frame control
Can interpolate motion between supplied first and final images in image-to-video workflows.
Editing endpoints
LTX-2.3 Pro supports retake, extend, reframe, and HDR conversion in addition to generation.
Open ecosystem
Provides base, distilled, and quantized checkpoints plus training code, ComfyUI nodes, LoRAs, reference workflows, and LTX Desktop.
Process
How the LTX-2.3 workflow works
Step 1
Choose local or API
Review the model license and available hardware before selecting self-hosted checkpoints or managed endpoints.
Step 2
Select Fast or Pro
Use Fast for lower-cost text and image iteration, or Pro when quality, audio conditioning, and editing endpoints matter.
Step 3
Set the controls
Provide the prompt and optional image, audio, last frame, camera motion, resolution, orientation, duration, and frame rate.
Step 4
Generate and finish
Review synchronization and continuity, retake weak sections, reframe if needed, then complete editing and disclosure requirements.
Cost
LTX-2.3 pricing and free plan
Self-hosted weights have no per-generation fee but require infrastructure and compliance with the LTX model license. Managed API generation is billed per second, with cost increasing by model, endpoint, and resolution.
Self-hosted LTX-2.3
No per-generation fee
Run the checkpoints locally or on owned infrastructure.
- Compute, storage, and engineering costs still apply
- Entities under $10M annual revenue can use the community license subject to its terms
- Larger commercial entities require a paid license
- Review redistribution and use restrictions
LTX-2.3 Fast API
$0.03-$0.24 per output second
Lower-cost text-to-video and image-to-video.
- 720p: $0.03/second
- 1080p: $0.06/second
- 1440p: $0.12/second
- 4K: $0.24/second
LTX-2.3 Pro API
$0.04-$0.32 per output second
Higher-quality text-to-video and image-to-video.
- 720p: $0.04/second
- 1080p: $0.08/second
- 1440p: $0.16/second
- 4K: $0.32/second
LTX-2.3 Pro audio-to-video
$0.06-$0.34 per input second
Generates video synchronized to supplied audio.
- 720p: $0.06/second
- 1080p: $0.10/second
- 1440p: $0.18/second
- 4K: $0.34/second
Editing and finishing APIs
$0.10-$0.80 per second
Retake, extend, reframe, and HDR operations have endpoint-specific rates.
- 1080p retake and extend: $0.10/second
- Reframe: $0.10 at 720p or $0.20 at 1080p
- HDR: $0.20 at up to 1080p, $0.40 at 1440p, $0.80 at 4K
- Billing basis differs by endpoint
Pricing checked . Check current pricing at the source ↗
Assessment
LTX-2.3 strengths and limitations
Where it stands out
- Generates synchronized audio and video within one open-weight model family
- Supports both local control and a managed per-second API
- Native portrait, first-to-last frame, retake, extend, reframe, and HDR make it useful beyond basic prompting
- Mature ecosystem of checkpoints, ComfyUI workflows, code, LoRAs, and desktop tooling
What to consider
- LTX-2.5 is now the newer family and may be the better default for new text, image, or audio generation integrations
- Local inference requires substantial compatible hardware and hands-on model operations
- The community license is not unrestricted open source; larger commercial entities need a paid license
- High-resolution or long API generations can become expensive because charges accrue per second
- Generated video, audio, faces, physics, and continuity still require human review and often multiple attempts
Compare
LTX-2.3 alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Content Creator
LTX-2.5
Choose LTX-2.5 for Lightricks' newer model family and the default direction for new integrations.
Explore LTX-2.5 →Content Creator
Runway Gen-4.5
Choose Runway Gen-4.5 for a managed cinematic model inside Runway's broader creative suite.
Explore Runway Gen-4.5 →Content Creator
Veo 3.1
Choose Veo 3.1 for Google's high-end video generation and native audio workflows.
Explore Veo 3.1 →Questions
LTX-2.3 FAQs
Is LTX-2.3 still available?
Yes. LTX-2.3 Fast and Pro remain supported in the LTX API and the checkpoints remain available for local use, although LTX-2.5 is now newer.
Is LTX-2.3 open source?
Lightricks releases weights, code, and workflows under its LTX community license. It is open-weight and modifiable, but the license includes use and commercial restrictions, including paid licensing for entities with at least $10 million in annual revenue.
Can LTX-2.3 generate vertical video?
Yes. It was trained for native portrait output and the API supports portrait resolutions from 720p through 4K.
Does LTX-2.3 generate audio?
Yes. The Pro model can jointly generate synchronized audio and video and can also create video driven by an input audio track.
How much does the LTX-2.3 API cost?
Text-to-video and image-to-video currently range from $0.03 to $0.24 per second on Fast and $0.04 to $0.32 on Pro, depending on resolution. Audio and editing endpoints use separate rates.
Bottom line
Our LTX-2.3 verdict
LTX-2.3 remains a capable and unusually flexible open-weight video engine, especially for teams that need synchronized audio, local deployment, native portrait, or the Pro editing endpoints. New projects should compare it directly with LTX-2.5 before locking the model choice, and every commercial self-hosted deployment should review the license rather than assuming the weights are unrestricted.
Visit LTX-2.3 website ↗Similar Content Creator Tools
View all Content Creator AI tools →
Nano Banana 2
Nano Banana 2 - Google's new No.1 ranked AI image model at Flash speed and pricing
Phota Studio
Phota Studio - Phota's personalized photo editing and generation model that preserves your identity across edits
Higgsfield Soul 2.0
Higgsfield Soul 2.0 - New creative model built for aesthetics, realism, and culture-native generations
Veo 3.1 Lite
Veo 3.1 Lite - Google's new budget-friendly video generation mode
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.