Kling AI releases 4.0 with stereo audio and multilingual lip sync

Kling AI 4.0 generates clips up to 30 seconds long and accepts up to ten keyframe images to lock transitions and character states, with prompt understanding expanded to 8,000 tokens. The upgraded Omni Reference takes up to 15 reference assets per generation, including 10 images, up to five videos totaling 30 seconds, and 7 elements, so storyboards, wireframes and depth-map videos can all serve as references. A new creative-replication feature lets users use an existing video as a template to swap in products or characters, adjusting expression, body movement, camera angle or background independently without affecting the rest of the frame. On audio, Kling 4.0 adds two-channel stereo, and lip sync covers German, English, Japanese, Korean, Spanish and more than a dozen other languages plus multiple dialect accents; in-video text generation is also multilingual and stays stable during camera motion. 4K and 1080p 10-bit HDR output and an extension feature that stretches clips to two minutes are both marked as coming soon. The companion lightweight Kling 4.0 Flash targets faster generation and lower cost but caps at 720p and 20 seconds. The creation interface has been redesigned, and Kling 4.0 now offers MCP and CLI access for building automated workflows via AI agents. External comparisons show Kling 4.0 trading wins with Wan 3.0 and Seedance 2.5 on shot logic, physical behavior and dialogue naturalness without a clear lead. API pricing is not yet announced and support is listed as coming soon, so developers using the Kling API cannot yet estimate integration costs.