Kling VIDEO 3.0 Omni
Video modelCard summarizedAll-in-one multimodal input, voice-driven characters, and native audio output with storyboards
Source: Kling VIDEO 3.0 Omni Model User Guide (official model guide; no system card published)
Where to call it
- Kling AI API (new API)
POST /omni-video/kling-3.0-omni
Good for
- Keeps characters, items and scenes consistent across shots from reference images or elements
- Combines multiple elements and reference images in group or interactive scenes
- Video-character elements capture appearance and the original voice from a 3–8 s clip
Overview
Kuaishou · 02Kling Omni, front
Kling 3.0 Omni
All-in-one multimodal input, voice-driven characters, and native audio output with storyboards
Typical price
$0.14 / sec
1080p, no video input, native audio
- Max resolution
- 4K
- Max length
- 15 s
- Modes
- text-to-video, image-to-video (first frame / first + last frame)
- Released
- 25 Feb 2026
VIDEO 3.0 Omni is the upgrade of Kling VIDEO O1, built on the Kling 3.0 unified multimodal training framework. It treats text, images, videos and elements as prompts, adds native audio and multi-shot storyboards, binds voices to character elements, and raises the single-generation limit from 10 s to 15 s. User guide, intro
kling-v3-omni- Typical price
- $0.14 / sec
- Max resolution
- 4K
- Max length
- 15 s
- Aspect ratios
- 16:9, 9:16, 1:1
- Frame rate
- 24
- Released
- 25 Feb 2026
- text-to-video
- image-to-video (first frame / first + last frame)
- multi-image reference (up to 7 images + elements)
- element reference (multi-image and video-character elements with bound voices)
- feature reference video
- video editing (base video)
- multi-shot storyboards (1–6 shots)