Kling 4.0 arrives on Higgsfield with clips up to 30 seconds and 10 keyframes

What Kling 4.0 offers

Higgsfield describes Kling 4.0 as Kling's latest video model, built for realistic motion, stable camera movement and improved visual and audio quality. According to the company, it handles fast action such as fights, sports and large-scale scenes, and keeps characters recognizable across scenes and variations.

The model runs on Higgsfield, which calls itself an AI-native creative suite with 35+ models and tools. Kling 4.0 sits there alongside Kling 3.0 and other video models.

Key settings on Higgsfield:

• Modes: text-to-video, image-to-video, first and last frames, multi-keyframes, Omni Reference and video editing
• Duration: 3 to 30 seconds
• Resolution: 720p and 1080p
• Aspect ratio: Auto, 21:9, 16:9, 9:16 and 1:1
• Prompt length: up to 8,000 tokens
• Audio: stereo with two channels

Changes compared with Kling 3.0

Higgsfield says Kling 4.0 doubles the maximum length of a generation, from 15 to 30 seconds. Keyframe control also grows: Kling 3.0 offered a first and last frame, while Kling 4.0 allows up to 10 keyframes across the video.

References expand from up to 7 in Kling 3.0 to up to 10 images and 5 videos in Kling 4.0, with up to 15 combined in one generation. The 21:9 aspect ratio is new, and voice support adds Portuguese, German, French, Hindi and more. Voices cover American, British, Indian and other English accents, as well as several Chinese dialects. On-screen text is supported in nine languages, plus emoji and logos.

For existing footage, Higgsfield lists integrated multimodal editing with more flexible reference inputs and targeted changes. The model also introduces Creative Recreation, which uses an existing video's tempo, composition and motion to guide a new scene. Video editing accepts up to 5 clips, one of them the main clip, each 3 to 30 seconds long, with a combined length of up to 30 seconds.

How the workflow and pricing work

Users open Video, then Create Video, and choose Kling 4.0 under Model. They then pick References, Frames or Edit video, add inputs and write a prompt. Higgsfield advises that when keyframes are used, they should set the key moments while the prompt describes the action and camera movement between them. If a result is off, the company suggests changing one element at a time, such as the prompt, a reference or the duration.

Generations cost credits, and the exact price appears on the Generate button before starting. Cost depends on resolution and clip length:

• 720p, 5 seconds: 25 credits, about $1.23
• 720p, 30 seconds: 150 credits, about $7.35
• 1080p, 5 seconds: 37.5 credits, about $1.84
• 1080p, 30 seconds: 225 credits, about $11.03

The dollar figures are approximations based on Higgsfield's mid-tier Plus plan, which costs $49 per month for 1,000 credits, or $0.049 per credit.

What it means for creators

Longer single clips mean actions and interactions have more time to develop before another generation is needed, according to Higgsfield. Keyframes and multiple reference types are meant to help keep characters consistent across scenes, while stereo audio and multilingual voices target dialogue-driven material. The editing tools let users change a character, product, background, camera angle or visual style in an existing clip.

Since cost scales with both length and resolution, a full 30-second clip at 1080p costs 225 credits, 45 times the 5-second 720p entry price of 25 credits only in the sense of length and resolution combined. Users can check the price before each generation.