Kling AI has launched version 4.0, shifting its focus from visual realism to enhanced directorial control. The update introduces tools for managing motion, character consistency, and narrative flow in longer video sequences.
The model now supports native generation of videos up to 30 seconds long. Users can also extend existing clips, facilitating more complex storytelling and continuous action sequences.
New features include support for up to 15 multimodal reference assets, such as images, videos, and audio. Additionally, users can utilize up to 10 keyframes to define specific character states, actions, and camera movements.
These tools aim to address industry challenges regarding narrative coherence and stability. By providing more precise inputs, the model seeks to reduce the need for repeated generations and manual trial-and-error.
