AivexaNewsSearch
AI news for builders and product teamsChecked every hour

Kling AI drops version 4.0 with stereo audio, multilingual lip sync, and a lighter Flash variant

Collected Oct 19, 2026

Kling AI has shipped version 4.0 of its AI video generator, adding longer single-pass clips, richer reference handling, stereo audio and multilingual lip sync, alongside a lighter Flash model built for faster and cheaper generation. The release targets higher resolution and more precise control, and the company frames it as narrowing the gap between generated and professional video.

The headline capability is duration: users can now produce up to 30 seconds of video in one pass. Up to ten keyframe images can be supplied as input to lock down scene transitions, character states and story beats, and prompt comprehension has been expanded to 8,000 tokens, which matters for creators who pack detailed shot descriptions, style notes and continuity instructions into a single prompt.

The upgraded Omni Reference system accepts up to 15 reference assets per generation, according to the company: as many as 10 images, 5 videos totalling at most 30 seconds, and 7 elements. Accepted reference formats include storyboards, wireframes, image grids, white-model (untextured) videos and depth-map videos. In practice, depth maps give a model explicit spatial structure, while white-model clips supply motion and geometry without surface detail, so a team can iterate on blocking and camera moves before committing to a look.

A Creative Recreation feature turns existing videos into templates, letting users swap products or characters and adjust individual elements such as facial expressions, body movements, camera angles or backgrounds without disturbing the rest of the clip. For product and marketing teams, that suggests a route to variant generation from one approved master shot, though the source does not detail how reliably elements stay isolated across edits.

On audio, Kling 4.0 adds two-channel stereo and improved lip sync across more than ten languages, including German, English, Japanese, Korean and Spanish, along with various dialects and accents. In-video text generation now works in several languages and holds up during camera movement, a detail that matters for signage, packaging and on-screen labels, which often warp when cameras pan.

Two features are listed as coming soon rather than available: 10-bit HDR output at 4K and 1080p, and a video extension function that would stretch clips to as long as two minutes. Until the extension ships, the 30-second single-pass limit stands as the practical ceiling, and longer sequences will require stitching.

The Flash variant trades fidelity for throughput and price. Kling 4.0 Flash is designed for faster generation at lower cost but caps out at 720p and 20 seconds. That is a familiar split in generative video: a quality tier for final output and a fast tier for ideation, prompt testing and storyboard drafts. A likely trade-off is that teams will prototype on Flash and finish on the full model, accepting some visual mismatch between the two.

The creation interface has been redesigned, and Kling 4.0 is reachable through MCP (the Model Context Protocol, an interface that lets AI agents call external tools) and a CLI, opening the door to automated pipelines driven by agents. Pricing for the Kling 4.0 API has not been announced and API support is listed as coming soon, so cost planning for programmatic use is not yet possible.

Independent comparison against Wan 3.0 and Seedance 2.5, both from Chinese competitors, shows Kling 4.0 holding its own without pulling ahead, particularly on logical shot sequencing, physics and natural-sounding dialogue. This suggests progress among leading video models may be converging, though the next Wan and Seedance releases will test that reading. In practice, buyers may increasingly choose on workflow features like reference formats, MCP access and cost rather than raw output quality alone.

Why it matters: developers building video pipelines can now script Kling 4.0 through MCP and a CLI, with longer clips, many more reference assets and multilingual lip sync to design around, while a 720p Flash tier at 20 seconds opens a cheap draft stage. The unannounced API price and the fact that HDR and two-minute extension remain forthcoming are the main planning unknowns; the comparison results suggest capability parity, so integration and workflow features are where teams will differentiate.

Read at The Decoder

Based on reporting from the original publisher. Visit the source for full context and later updates.

Publisher excerpt

Kling AI has released version 4.0 of its video generator with support for clips up to 30 seconds, ten keyframe images as input, and prompt comprehension expanded to 8,000 tokens. In comparisons with Wan 3.0 and Seedance 2.5, Kling 4.0 holds its own but doesn't pull ahead, which could be an early sign that the breakneck pace of progress in AI video models is starting to level off. The article Kling AI drops version 4.0 with stereo audio, multilingual lip sync, and a lighter Flash variant appeared first on The Decode