LTX 2.5 Pro

LTX 2.5 Pro — Image-to-Video and Text-to-Video with Native Audio

What is LTX 2.5 Pro?

LTX 2.5 Pro is the high-fidelity tier of Lightricks' LTX 2.5 video family, generating video with synchronized audio in a single pass. It handles both text-to-video and image-to-video, producing portrait or landscape clips up to 1080p at cinematic frame rates. Built on a 22-billion-parameter diffusion transformer, it pairs a new Diffusion Video Decoder with Diffusion Fidelity Rendering that allocates compute by scene complexity, so faces, textures, and fast motion hold up frame by frame. Its standout capability is native multi-shot generation: a single prompt can render several connected shots that keep character, environment, lighting, and voice consistent across cuts.

Key Features

• •Text-to-video and image-to-video from one endpoint, up to 1080p in 16:9 or 9:16.
• •Native synchronized audio (ambient sound, music, and dialogue) generated jointly with video.
• •Native multi-shot scenes with consistent characters and voice across explicit cuts.
• •Diffusion Video Decoder and Diffusion Fidelity Rendering for cleaner high-motion output.
• •Automatic duration, camera-motion controls, and first-frame plus last-frame inputs.

Best Use Cases

LTX 2.5 Pro is built for production-quality work: concept films, storyboards, product demos, social ads, and speaking-character shots. Its multi-shot consistency suits branded, narrative content where a character and setting must survive cuts. The image-to-video path animates an approved still with matching audio when you need a locked look. Landscape 16:9 fits cinematic edits, while 9:16 targets vertical social formats.

Prompt Tips and Output Quality

Write one flowing, present-tense paragraph covering shot, scene, action, character, camera, and audio. Put camera moves in their own clause, describe sound explicitly, and place spoken dialogue in quotation marks. Keep on-screen text short and add critical titles in post, since exact spelling across frames is not guaranteed. For multi-shot prompts, name each cut and state audio continuity.

FAQs

Does LTX 2.5 Pro generate audio? Yes, it creates a synchronized track jointly with the video; set generate_audio to false for silent output.

What resolutions and durations are supported? Up to 1080p (720p also) in 16:9 or 9:16, at 6, 8, or 10 seconds and 24, 25, or 50 fps.

Does it support image-to-video? Yes. Supply a first-frame image to switch modes, and optionally a last frame to interpolate toward.

What is native multi-shot generation? One prompt can render multiple connected shots that keep character, lighting, and voice consistent across cuts.

How is Pro different from LTX 2.5 Fast? Pro targets higher fidelity up to 1080p, while Fast trades some quality for speed and reaches up to 4K.