The AI video race has a new headline number: 30 seconds. Kling AI, the creative platform built by Chinese technology company Kuaishou, announced the global launch of Kling 4.0, its latest video generation model, in a GlobeNewswire release that has been rolling out to users through this weekend.
The upgrade works on three fronts, the company says: more stable dynamic motion and complex camera movements; native 30-second generation for long takes and continuous storytelling — double the previous generation's 15-second ceiling; and an enhanced Omni Reference system supporting up to 15 multimodal reference assets and 10 keyframe inputs, so creators can hold characters, scenes and story beats consistent across shots.
Kling says the platform now serves more than 100 million users globally and more than 50,000 enterprise customers across advertising, film, television and gaming. Support for 10-bit HDR output at 4K and 1080p, and a 21:9 ultrawide format, is listed as coming soon rather than available at launch — a distinction honest reviewers of the release have been careful to keep.
For working creators, the significance is less any single specification than the direction of travel: longer single takes, keyframe-level planning and larger reference budgets attack exactly the continuity problems that have kept AI clips hard to use in professional production. Kling 4.0 arrives in a three-way contest with Google's Veo and ByteDance's Seedance for the tools filmmakers actually reach for.
