Kling AI has launched version 4.0 of its video generation model, introducing higher resolution capabilities and more precise control tools. The update includes a new Flash variant designed for faster, lower-cost generation, which supports up to 720p resolution and 20-second clips.
The platform now allows users to generate up to 30 seconds of video in one pass and supports prompt comprehension of up to 8,000 tokens. Users can input up to ten keyframe images to manage scene transitions and character states, while the Omni Reference system accepts up to 15 assets, including images and videos.
New features include Creative Recreation for modifying specific elements like backgrounds or camera angles in existing videos. The model also adds two-channel stereo audio and improved lip-sync support for over ten languages, including English, German, Japanese, Korean, and Spanish.
The interface has been redesigned to support access via MCP and CLI for automated workflows. While 4K and 1080p 10-bit HDR output and a two-minute video extension function are listed as upcoming features, pricing for the Kling 4.0 API has not yet been disclosed.
