Kling AI launches version 3.0 with 15-second videos and multilingual audio

The new Kling AI 3.0 model series includes four models: Video 3.0, Video 3.0 Omni, Image 3.0, and Image 3.0 Omni, representing a significant advance in narrative control and visual coherence.

Video 3.0 incorporates native audio generation in English, Chinese, Japanese, Korean, Spanish, and various accents and dialects. This enables the creation of complex dialogue scenes between multiple characters, each speaking a different language. The maximum video duration extends to 15 seconds, sufficient for elaborate sequences with multiple narrative twists and cinematic transitions.

Among the standout improvements is visual element consistency. Creators can upload reference videos and multiple images to ensure characters, objects, and scenarios maintain coherence across frames. The model understands multi-scene and multi-shot instructions, dynamically adjusting camera angles according to creative direction.

The system also improves text preservation in images, maintaining signage, subtitles, and brand elements with high precision. This capability proves useful in e-commerce advertising, where logos on clothing remain sharp throughout the entire video.

Video 3.0 Omni expands reference capabilities by allowing the AI to extract visual traits and voice characteristics from a character to replicate them in new scenes. It incorporates a multi-shot storyboard feature where users specify duration, framing, perspective, and camera movements for each shot.

The image models Image 3.0 and Image 3.0 Omni support 2K and 4K output for professional use cases, preserving textures, lighting, and material qualities with notable precision.

Since its launch in June 2024, Kling AI has over 60 million creators globally and has produced more than 600 million videos. The models are available in early access for Ultra subscribers and will soon open to the general public.

Key points

• Kling AI launches four new models: Video 3.0, Video 3.0 Omni, Image 3.0, and Image 3.0 Omni
• Videos reach up to 15 seconds with cinematic transitions
• Native audio in English, Chinese, Japanese, Korean, Spanish, and various accents
• Enables dialogues between characters speaking different languages
• Improves visual consistency with reference videos and images
• Preserves text in images with high precision for advertising
• Includes multi-shot storyboard with framing and movement control
• Images in 2K and 4K resolution for professional use
• Over 60 million creators and 600 million videos generated