Kling 4.0

• 🆕 Kuaishou's fourth-generation Kling video model, image-to-video endpoint
• 👁️ Animates a starting frame, with optional last frame and keyframes
• 📏 Clips up to roughly 30 seconds in one generation
• 🎯 Up to 10 keyframes to pin story beats precisely
• 🔧 Omni Reference accepts image, video, subject and voice inputs
• 💬 Synchronized audio generated alongside the picture, per Kling
• 🌐 720p and 1080p output, with 21:9 ultrawide framing
• 🧠 Kling describes steadier complex motion and better camera-sequence understanding

Kuaishou

Kuaishou Technology is a Chinese publicly traded company founded in 2011 by Hua Su and Cheng Yixiao, headquartered in Beijing's Haidian District and listed on the Hong Kong Stock Exchange. Originally known for its massively popular short-video platform —…

Kling 4.0 is the image-to-video entry point into Kuaishou's fourth-generation video model, announced by Kling AI in a dedicated release note on its own site. You supply a starting frame (optionally a final frame and intermediate keyframes) plus a prompt, and the model animates the scene with synchronized audio generated alongside the picture. It sits beside Kling 4.0 text-to-video and Kling 4.0 R2V in the same release.

Against the earlier V3-era models — Kling V3 Pro, Kling V3 Turbo Pro and the reasoning-oriented Kling O3 Pro — Kuaishou positions 4.0 as a native multimodal model with more stable complex motion, stronger consistency across diverse references, and more coherent, controllable storytelling. Kling's own notes also highlight a stronger grasp of camera movement and shot sequencing, better highlight and shadow retention in backlit, night and neon scenes, less color banding, and new 21:9 ultrawide framing.

Practically, the generation window stretches to about 30 seconds in one take, with up to 10 keyframes for beat-by-beat control and an Omni Reference input that accepts images, video clips, subjects and voice. Reference video inputs can include white-model and depth-map footage to steer motion, camera and timing.

The image-to-video route takes a still image plus a text prompt, optional end frame, keyframes and references, and returns a video clip with audio. Older entries such as Kling 2.6 Pro remain available for shorter generations.