Kling
Follow Kling from a limited text-to-video beta to native sound, reusable references, in-video editing, and controllable multi-shot production across Kling AI, Runway, and Higgsfield.
- Posts archived
- Attachments examined
- Capabilities with evidence
What changed, when
Announcement dates in KST · Select a change or open its example.
Workflow threads
Compare workflows
Compare product availability and limits
Shots
- Kling AI · · Kling 4.0 adds keyframes within a clip of up to 30 seconds.
- Runway · · The API exposes first, middle, and last frame inputs plus mixed references.
- Higgsfield · · Frames mode allows up to ten keyframes in one clip.
References
- Kling AI · · Motion Control borrows action from a reference video.
- Runway · · Frame and media references are separate input routes.
- Higgsfield · · References can combine images, videos, and saved Elements.
Edit
- Kling AI · · Kling 4.0 integrates multimodal editing of an existing clip.
- Runway · · Runway Dev exposes a source-video edit route.
- Higgsfield · · Edit video covers character, product, backdrop, and style changes.
Audio
Workflows over time
Announcement dates in KST. Earlier notices establish what was already available.
Update archive
Official titles with edited summaries and post-level sources.
Official examples and illustrations
Play video in its card or enlarge the example for its source and context.
Read the dated updates and sources
- · Kling opens limited video testing
Kuaishou began accepting applications to test text-driven video generation in its KuaiYing video editor in China. The June 10 press release describes the limited beta. Kuaishou dates first applications to June 6 in its global beta notice.
- · Kling opens its web beta worldwide
The upgraded beta provides text- and image-to-video, clip extension, and improved motion and visual quality. The release text places the announcement on July 25. The investor page header displays July 26. Subscription benefits listed in the release apply to mainland China.
- · Kling moves clip extension into its video workflow
Kling 1.6 adds extension in five-second steps. A separate AI Sounds tool can generate sound from a prompt or using a video as input. The release note describes a maximum of three minutes with extension.
- · Kling 2.6 generates audio with the picture
Create a clip with dialogue, action sounds, and environmental ambience generated together with the visuals. Kuaishou's December 5 investor announcement identifies December 3 as the launch date.
- · Kling 3.0 adds shot-level direction
The video series adds multi-shot direction, subject references, and speech across multiple languages. Omni combines image, video, and audio inputs for generation and editing. Kling's release index labels the Video 3.0 item January 31. Kuaishou published its investor announcement February 5.
- · Higgsfield separates Kling generation and Edit
The model selector offers standard generation, multimodal Omni, and a dedicated Omni edit route for existing clips. Standard Kling 3.0 clips are 3–15 seconds at 720p or 1080p in the product note.
- · Kling 3.0 upgrades motion transfer
Drive a character from a motion-reference video, with improvements in facial consistency and movement precision. The motion-control variant has its own input and orientation rules.
- · Kling previews a 30-second, multi-keyframe workflow
The preview demonstrates longer native generation, up to ten keyframes, reference assets, restaging an existing clip, and richer audio controls. The September announcement is a preview. Kling dates the full 4.0 launch to October 15.
- · Runway Dev adds Kling 2.5 and 3.0 variants
Runway opens Kling 2.5 Turbo Pro and the 3.0 Standard, Pro, 4K, Motion Control, and Omni routes. Endpoint choices determine available frame inputs, resolution, and audio. The API changelog is the evidence for this surface. Its model IDs and limits vary by variant.
- · Kling 4.0 launches with up to ten keyframes
Set key moments inside a clip, guide generation with mixed image and video references, and generate up to 30 seconds with sound. Kuaishou confirms October 15 as the launch date. The September preview, October launch, and downstream adoptions are separate milestones.
- · Higgsfield exposes References, Frames, and Edit
Use up to ten keyframes and up to 15 combined references, or swap elements in an existing video. The notes list up to 30 seconds and stereo sound. The product lists 720p and 1080p and supports a 3–30 second duration range.
- · Runway Dev adds Kling 4.0 generation and editing
Use text, frames, mixed image and video references, or a source video. The API supports 3–30 seconds at 720p or 1080p. The default is five seconds at 720p. Check the endpoint for input-specific rules.
- · Higgsfield adds Kling 4.0 Flash for drafts
Flash supports References, Frames, and Edit video at 720p for 3–20 seconds. On the Frames route the start image sets the aspect ratio. The product notice specifies a start frame only for Flash.
- · Runway Dev adds Kling 4.0 Flash
Generate 3–20 seconds of 720p video with audio from text, a first frame, or media references. A source-video route supports editing. The API model ID is kling_4_flash. Each input category has its own count or duration limit.
Documentation and availability notes
- KuaiYing · · The June 10 notice describes trial enrollment in China. Its maximum model length is a capability claim, and practical duration depends on the interface.
- Kling AI · · The press announcement is dated February 5 and follows Kling's January 31 release-index entry and the February 4 Higgsfield listing.
- Kling API · Use the capability map for current output sizes, reference types, and aspect ratios per variant.