

ltx 2.3 is Lightricks' video generation family for cinematic creation across Fast and Pro workflows. Fast models can produce clips up to 20 seconds at 720p or 1080p, while image workflows can animate a first frame and optionally interpolate toward a final frame. Access the family through one OpenAI-compatible Atlas Cloud API with transparent per-call pricing. Start building today.
Atlas Cloud (operated by Atlas Cloud AI LLC) provides access to third-party AI models developed and owned by their respective providers. Atlas Cloud does not own these models, and all model names, product names, and trademarks belong to their respective owners.
Atlas Cloud provides you with the latest industry-leading creative models.
Explore each LTX 2.3 modality by input type, output format, generation mode, resolution, frame rate, duration, and audio support.
| Modality | Description |
|---|---|
| LTX 2.3 Pro API (Audio to Video) | Transform dialogue, music, or ambient audio into video synchronized with the driving track, with the source audio determining the duration. Generate at up to 4K and 24 or 48 fps for music visuals, dialogue-led scenes, and audio-first productions. |
| LTX 2.3 Fast API (Image to Video) | Animate a first-frame image into silent video, with optional interpolation toward a supplied last frame. Supporting up to 4K at 24 or 48 fps and up to 20 seconds at 720p or 1080p, this endpoint suits rapid motion concepts and visual prototypes. |
| LTX 2.3 Fast API (Text to Video) | Describe a scene in text and receive a silent cinematic clip generated at up to 4K and 24 or 48 fps. Clips can reach 20 seconds at 720p or 1080p, making the endpoint practical for fast concept visualization and content iteration. |
| LTX 2.3 Pro API (Image to Video) | Starting from a first-frame image, this endpoint creates a 6 to 10 second video with synchronized audio and can optionally interpolate toward a last frame. Output reaches up to 4K at 24 or 48 fps for polished image-led sequences. |
| LTX 2.3 Pro API (Text to Video) | Create cinematic video with synchronized audio directly from a text prompt. With 6 to 10 second output, resolution up to 4K, and 24 or 48 fps support, it fits narrative concepts, campaign scenes, and audiovisual story development. |
| LTX 2.3 Quality API (Text to Video) | The Quality text-to-video endpoint converts written prompts into high-quality LTX 2.3 video with audio. Choose it for prompt-driven productions where refined audiovisual output matters more than the faster generation path. |
| LTX 2.3 Quality API (Image to Video) | Bring source images into motion as high-quality LTX 2.3 video with audio. This modality is suited to image-based storytelling, animated campaign assets, and projects that begin with an established visual composition. |
| LTX 2.3 Quality API (Extend Video) | Continue an existing clip through the dedicated LTX 2.3 Quality video-extension endpoint. Use this option when a production needs additional footage beyond its original video rather than a new sequence generated from scratch. |
LTX 2.3 combines synchronized audio generation, audio driven video, text and image workflows, 4K output, flexible frame rates, extended Fast clips, and guided transitions through one Atlas Cloud API.
LTX 2.3 Pro generates video and synchronized audio together from text or image inputs. Its 6 to 10 second clips support dialogue, music, and ambient sound while reaching up to 4K at 24 or 48 fps. Choose it when sound belongs inside the scene from the first render. The result suits cinematic ads, music moments, and narrative product films.
Start with dialogue, music, or ambience and let the Pro audio to video endpoint build synchronized visuals around that track. The source audio determines the output length, while generation can reach 4K at 24 or 48 fps. A prompt can guide the scene as the soundtrack drives timing and performance. This workflow fits performances, lyric visuals, and atmosphere led storytelling.
Resolution and motion are configurable across the family, with supported output up to 4K and frame rates of 24 or 48 fps. Use the higher ceiling for detailed final renders, or select a lighter workflow when iteration speed matters. These options help teams shape footage for cinematic displays, polished campaigns, and motion heavy scenes without changing model families.
When iteration speed matters, the Fast text to video and image to video endpoints create silent clips up to 20 seconds at 720p or 1080p. They also support output up to 4K at 24 or 48 fps for shorter configurations. Explore motion, pacing, and composition before choosing a Pro render with synchronized audio. It is a practical route for storyboards, concept reels, and rapid creative testing.
Provide a first frame to establish the opening composition, then describe the motion that should unfold. For more directed transitions, supported image to video workflows can interpolate toward an optional last frame. This gives a project clear visual anchors at both ends while leaving room for generated movement between them. Use it for transformations, product reveals, and planned scene transitions.
The LTX 2.3 family covers text to video, image to video, audio to video, and video extension through dedicated Atlas Cloud endpoints. Choose Fast for silent iteration, Pro for 6 to 10 second synchronized audio clips, or Quality when an extension workflow is required. A unified API keeps these routes together with transparent pay as you go base pricing. This range supports both experimentation and production handoffs.
See how LTX 2.3 and two alternative video models interpret identical prompts across live action and stylized motion challenges.
An 8–10 second continuous micro-adventure inside a decaying deep-sea shipwreck: a clever coconut octopus magician, with consistent mottled amber skin and expressive eyes, uses all eight arms independently and simultaneously—carrying luminous warm-gold pearls, sweeping aside swaying seaweed, spinning a corroded brass key inside an ancient lock, gripping the hull, and balancing a drifting dinner plate. Begin with an extreme macro tracking shot gliding across flexing suckers as each arm performs its task with precise coordination. The cabin door suddenly collapses with a heavy underwater crash, scattering rust flakes, bubbles, silt, and splintered wood; the octopus compresses its soft body, blasts a vivid purple ink cloud, and jets backward. Whip-pan along the curling ink trail as it strikes the floating plate, rebounds elastically, reshapes its body, and shoots cleanly through a narrow porthole while loose pearls and debris collide with believable underwater momentum. Transition into a fast orbit around the porthole, following it as it unlocks and opens a weathered shell; the largest pearl cracks and hatches into a tiny glowing squid. End from the newborn squid’s first-person POV as its translucent tentacles enter frame and the delighted octopus reaches toward it through sparkling particles. Cold cyan abyss pierced by narrow flashlight-like beams, warm golden pearls and rich ink-purple accents, volumetric caustics, realistic buoyancy, fluid dynamics, soft-body deformation, object impacts, continuous energetic motion, premium marine nature documentary cinematography with subtle photorealistic fantasy, crisp tactile detail, no text, no interface, no diagrams. Immersive audio: muffled hull groans, lock clicks, shell crack, bubbling jets, rushing water, metallic plate clang, and a delicate magical chime at the hatchling reveal; tense rhythmic pulse resolving into wonder. 16:9 aspect ratio.
Generated with LTX-2.3 Pro Text-to-Video on Atlas Cloud
Generated with Kling V3.0 Turbo Text-to-Video on Atlas Cloud
Generated with LTX-2.3 Fast Text-to-Video on Atlas Cloud
An 8–10 second continuous animated action sequence: above a raging ocean of thunderclouds, a teenage girl with brass goggles rides astride a colossal whale-shaped silk kite. Begin with an extreme low-angle tracking shot skimming inches above the violently buckling wing as lightning severs the tail rope; she braces against the gale, crawls rapidly across the pitching silk, and stitches a widening tear with glowing vermilion thread while loose ropes whip, tighten, and react with convincing tension. The kite suddenly folds into a steep dive; the camera accelerates into a full orbit and rolls upside down around her as she leaps, swings from the wooden thread spool, and threads through a circular opening in the clouds, preserving clear spatial continuity and consistent character design. Whip-pan into a high aerial perspective as she lands, yanks the final stitch tight, and the kite snaps open; pull rapidly backward during the climactic dive as its great whale tail lashes sideways, shattering the thunderheads into thousands of luminous silver fish that scatter across the stormy sky. Maintain dynamic diagonal compositions and continuous, fluid motion from first frame to last, with physically convincing silk deformation, rope dynamics, wind resistance, rain, lightning, and volumetric cloud turbulence. Hand-drawn cel animation fused with sculptural three-dimensional ink-wash clouds, cool cyan thunderlight contrasted by vivid scarlet stitching, crisp expressive linework, dramatic atmospheric depth, no slow motion. Synchronized audio: roaring wind, cracking silk, taut rope snaps, thunder impacts, rapid stitching sounds, a rising percussive score, and a bright metallic shimmer as the silver fish burst outward. No screens, software interfaces, dashboards, progress bars, charts, captions, subtitles, logos, watermarks, or explanatory text. 16:9 aspect ratio.
Generated with LTX-2.3 Pro Text-to-Video on Atlas Cloud
Generated with Kling V3.0 Turbo Text-to-Video on Atlas Cloud
Generated with LTX-2.3 Fast Text-to-Video on Atlas Cloud
From cinematic prompting and guided image animation to audio-driven visuals and rapid concept testing, LTX 2.3 supports varied production workflows through Pro and Fast video endpoints.
Turn written prompts into cinematic clips with LTX 2.3 Pro or Fast, supporting up to 4K at 24 or 48 fps. Product teams can produce campaign concepts, mood pieces, and short narrative scenes.
Animate a first-frame image, then optionally guide the ending with a last frame using LTX 2.3 image-to-video. Designers can turn key art, product stills, and storyboards into controlled motion sequences.
Provide dialogue, music, or ambience to LTX 2.3 Pro and let the track determine the video's duration while visuals stay synchronized. Creators can build music visuals, performance clips, and atmosphere-led sequences from existing sound.
Start from one image and interpolate toward a chosen last frame, with output available up to 4K at 24 or 48 fps. Creative teams gain clearer control over reveals, transformations, and scene transitions.
Choose Fast text-to-video or image-to-video for silent clips lasting up to 20 seconds at 720p or 1080p, with support up to 4K. Use these outputs for rapid concept testing, pitch visuals, and preproduction reviews.
Need a compact campaign asset? Generate 6 to 10 second Pro clips from text or a first-frame image, combining up to 4K visuals with synchronized audio for ads, launches, and social storytelling.
Compare LTX 2.3 variants with other available video models across duration, resolution, frame rate, synchronized audio, and standard base pricing.
| Model | Output Duration | Max Resolution | Frame Rate | Synchronized Audio | Base Price |
|---|---|---|---|---|---|
| LTX-2.3 Pro Text-to-Video | 6 to 10s | 4K | 24 or 48 fps | √ | $0.04 |
| LTX-2.3 Fast Text-to-Video | Up to 20s | 4K | 24 or 48 fps | - | $0.03 |
| Seedance 2.0 Text-to-Video | 4 to 15s | 4K | - | √ | $0.112 |
| Veo 3.1 Lite Text-to-video | 4, 6, or 8s | 1080p | 24 fps | √ | $0.05 |
Get started in minutes — follow these simple steps to integrate and deploy models through Atlas Cloud's platform.
Sign up at atlascloud.ai and complete verification. New users receive free credits to explore the platform and test models.
Combining the advanced ltx 2.3 First-to-Last Frame Animation models with Atlas Cloud's GPU-accelerated platform provides unmatched performance, scalability, and developer experience.
Low Latency:
GPU-optimized inference for real-time reasoning.
Unified API:
Run ltx 2.3 First-to-Last Frame Animation, GPT, Gemini, and DeepSeek with one integration.
Transparent Pricing:
Predictable per-token billing with serverless options.
Developer Experience:
SDKs, analytics, fine-tuning tools, and templates.
Reliability:
99.99% uptime, RBAC, and compliance-ready logging.
Security & Compliance:
SOC 2 Type II, HIPAA alignment, data sovereignty in US.
LTX 2.3 is Lightricks' model family for generating video from text, images, or audio. Atlas Cloud provides hosted Fast, Pro, and Quality endpoints, removing the need to deploy the underlying models yourself.
Create cinematic clips from prompts, animate a starting image, synchronize video with a driving audio track, or extend an existing video. The available workflow depends on the selected Fast, Pro, or Quality endpoint.
Create an Atlas Cloud API key and submit a request to the video generation endpoint with the required model ID and inputs. The API returns a prediction identifier that your application can poll until the generated output is ready.
Choose Fast for lower cost, silent text-to-video or image-to-video generation. Use Pro when your workflow needs synchronized audio or audio-driven video, while the Quality collection covers text-to-video, image-to-video, and video extension.
Fast and Pro endpoints support output up to 4K at 24 or 48 fps. Fast can generate up to 20 seconds at 720p or 1080p, while Pro text-to-video and image-to-video clips run from 6 to 10 seconds.
Yes, Pro text-to-video and image-to-video endpoints generate synchronized audio, and Pro audio-to-video follows the duration of the driving audio track. Fast text-to-video and image-to-video outputs are silent.
Image-to-video endpoints animate a supplied first frame. You may also provide a last frame on supported Fast and Pro workflows, allowing the model to interpolate the motion between the two images.
Standard pay-as-you-go pricing starts at $0.002 per second for Quality endpoints. Fast costs $0.03 per second, Pro text-to-video and image-to-video cost $0.04 per second, and Pro audio-to-video costs $0.06 per second.
Confirm that the model ID matches the intended workflow and that every required input follows that endpoint's schema. For image, audio, or video inputs, also verify that the supplied file URL is valid and accessible before resubmitting the request.
Guides, tutorials, and product updates to help you get the most out of Atlas Cloud.