On December 1, 2025, Kling AI made a significant announcement—the full - scale rollout of its brand - new video O1 model. This model has earned the title of the world's first unified multimodal video large model. It brings in the MVL (Multimodal Visual Language) interaction architecture and incorporates Chain - of - Thought technology. This innovative combination smashes through the barriers set by traditional video generation tools, which are often plagued by fragmented functions and cumbersome operations.

For users, the experience becomes incredibly streamlined. They can effortlessly blend various types of instructions—text, images, and videos—all within a single input box. This enables one - stop processing for a range of tasks, such as converting text into video, transforming images into video, conducting partial editing, and extending shots.

The model is equipped with robust common - sense reasoning and event deduction abilities. It can accurately grasp user intentions and generate video content that is more logically coherent. Take an example: after uploading a real - life video, users can simply issue commands in a conversational manner. They can add or remove elements locally, intelligently extend the preceding and subsequent shots, or capture actions to create new video footage.

Moreover, thanks to the multi - view subject construction technology, the O1 model completely overcomes an industry - wide problem. During shot transitions, characters or objects often suffer from feature drift, which can disrupt the visual continuity. The O1 model ensures that in multi - subject scenes, the visuals remain precise and coherent.

The model offers great flexibility in terms of generation duration, allowing free generation from 3 to 10 seconds. This puts the control of narrative pacing back into the hands of creators. Whether it's short video bloggers looking to create engaging content, advertising teams aiming for high - impact campaigns, or individual users with creative ideas, the Kling O1 model empowers them to rapidly produce high - quality, highly consistent creative videos.

At present, users can experience the Kling O1 model on the Kling App and its official website. There are also plans to open up API access in the future, enabling third - party platforms to integrate this powerful tool.