Kling AI drops version 4.0 with stereo audio, multilingual lip sync, and a lighter Flash variant
Kling AI has released version 4.0 of its video generation model, adding a cluster of features that expand both what the model can produce and how precisely users can direct it. Clips can now run up to 30 seconds, prompt comprehension has been extended to 8,000 tokens, and users can supply as many as ten keyframe images to guide a video's visual progression. Alongside the flagship update, the company is also shipping a lighter Flash variant aimed at faster generation with a reduced computational footprint.
On the audio side, version 4.0 introduces stereo sound and multilingual lip sync - meaning generated or edited video can now match spoken dialogue across multiple languages with synchronized mouth movement. These additions push Kling closer to being a more complete production tool rather than a purely visual generator, a direction several competing platforms have also been pursuing.
In side-by-side comparisons with Wan 3.0 and Seedance 2.5, Kling 4.0 performs competitively but does not establish a clear lead. That result is worth noting: for much of the past two years, new model versions reliably leapfrogged their predecessors and rivals by a visible margin. A tighter clustering of results across top-tier models may indicate that the steepest part of the quality improvement curve is beginning to flatten, at least temporarily, as the field matures and the low-hanging technical gains become harder to find.
For existing Kling users, the practical upgrades - longer clips, richer prompts, more keyframe control - are meaningful additions that extend what the tool can handle in a single generation pass. The Flash variant also signals that Kling's developers are thinking about the full range of use cases, from high-fidelity output to faster, iterative workflows where turnaround time matters more than peak quality. How these features hold up in real production environments will become clearer as more creators put version 4.0 through its paces.
