Kling O1 Standard (December)

Kling O1 Standard is a video generation model developed by Kuaishou's KlingAI division as part of its "Omni One" series. Launched in December 2025, it is a unified multimodal video engine designed to integrate text, image, and video inputs into a single creative workflow. The model is distinguished by its use of a Chain of Thought (CoT) reasoning system, which analyzes instructions before generation to improve motion accuracy, physical consistency, and prompt adherence.

Key capabilities of the model include Multi-Elements editing, which allows users to perform video-to-video transformations such as swapping objects, adding elements, or removing distractors from existing footage using natural language commands. It supports a wide range of tasks, including text-to-video generation, image-to-video conversion, and complex multi-reference processing using up to 10 images to maintain character and scene consistency.

As the Standard tier of the O1 release, the model focuses on efficient high-quality output, supporting video lengths of up to 10 seconds and resolutions up to 1080p. It operates on a Multimodal Visual Language (MVL) framework, enabling it to interpret complex director-level instructions regarding camera movement, lighting, and environmental changes without requiring manual masking or keyframing.