lightricks/ltx-video

OfficialView on Replicate →

LTX-Video is the first DiT-based video generation model capable of generating high-quality videos in real-time. It produces 24 FPS videos at a 768x512 resolution faster than they can be watched.

Capabilities

Reference ImagesNegative PromptSeed

Cost

Community model (estimated from hardware time)

Input Parameters

Name	Type	Description	Default	Constraints
`aspect_ratio`	string	Aspect ratio of the output video. Ignored if an image is provided.	`"3:2"`	1:11:22:12:33:23:44:34:55:49:1616:99:2121:9
`cfg`	number	How strongly the video follows the prompt	`3`	min: 1, max: 20
`image`	string(uri)	Optional input image to use as the starting frame	`—`	—
`image_noise_scale`	number	Lower numbers stick more closely to the input image	`0.15`	min: 0, max: 1
`length`	integer	Length of the output video in frames	`97`	97129161193225257
`model`	string	Model version to use	`"0.9.1"`	0.9.10.9
`negative_prompt`	string	Things you do not want to see in your video	`"low quality, worst quality, deformed, distorted"`	—
`prompt`	string	Text prompt for the video. This model needs long descriptive prompts, if the prompt is too short the quality won't be good.	`"best quality, 4k, HDR, a tracking shot of a beautiful scene"`	—
`seed`	integer	Set a seed for reproducibility. Random by default.	`—`	—
`steps`	integer	Number of steps	`30`	min: 1, max: 50
`target_size`	integer	Target size for the output video	`640`	5125766407047688328969601024

aspect_ratiostring

Aspect ratio of the output video. Ignored if an image is provided.

Default: "3:2"

1:11:22:12:33:23:44:34:55:49:1616:99:2121:9

cfgnumber

How strongly the video follows the prompt

Default: 3min: 1, max: 20

imagestring

Optional input image to use as the starting frame

image_noise_scalenumber

Lower numbers stick more closely to the input image

Default: 0.15min: 0, max: 1

lengthinteger

Length of the output video in frames

Default: 97

97129161193225257

modelstring

Model version to use

Default: "0.9.1"

0.9.10.9

negative_promptstring

Things you do not want to see in your video

Default: "low quality, worst quality, deformed, distorted"

promptstring

Text prompt for the video. This model needs long descriptive prompts, if the prompt is too short the quality won't be good.

Default: "best quality, 4k, HDR, a tracking shot of a beautiful scene"

seedinteger

Set a seed for reproducibility. Random by default.

stepsinteger

Number of steps

Default: 30min: 1, max: 50

target_sizeinteger

Target size for the output video

Default: 640

5125766407047688328969601024

Version: 8c47da666861Updated: 7/25/2026165.3K runs