Kling O1

Kling O1 is Kuaishou's first unified multimodal video model, making it exceptionally good at native video-to-video editing. Instead of just generating clips from scratch, it allows you to modify existing footage using natural language prompts—such as swapping backgrounds, changing a character's clothing, or shifting the lighting from day to night—without complex masking. Because of its fast generation speed and high prompt adherence, the community has dubbed it the "Nano Banana of AI video."

Kuaishou officially launched Kling O1 on December 1, 2025, kicking off its "Omni" lineup of unified multimodal models. It arrived shortly after the 2.x generation, such as Kling 2.6 Pro, and represented a major architectural shift from pure generation to a hybrid generation-and-editing engine. It was eventually succeeded by the 3.0 generation in February 2026, which includes standard models like Kling V3 Pro and the next-generation omni model, Kling O3 Pro.

To maintain strict character consistency while modifying footage, leverage Kling O1's multi-element reference capabilities by uploading reference images alongside your video. When prompting, use a conversational approach rather than traditional keyword stuffing. Because the model processes text, image, and video in a shared semantic space, clear instructions like "remove all cars from the street" or "change the protagonist's outfit to a red dress" yield much better results than comma-separated tags.