Kling 2.0 is a high-fidelity video generation model developed by KlingAI (Kuaishou Technology), succeeding the Kling 1.x series. Launched as a significant foundational upgrade, it emphasizes improvements in motion dynamics, prompt adherence, and visual aesthetics. The model is built on a Diffusion Transformer (DiT) architecture, enabling the generation of cinematic videos with enhanced temporal coherence and realistic physics.
Key Features and Control
A major innovation in Kling 2.0 is the Multi-modal Visual Language (MVL) system. This interactive approach allows users to provide multi-dimensional creative inputs—such as identity, style, and camera movement—by combining text prompts with multimodal references like image and video clips. It also introduced the Multi-Elements Editor, which allows users to swap, add, or delete specific objects within a video through simple text or image inputs, offering high flexibility for professional editing.
Technical Specifications
Kling 2.0 supports high-definition output at 1080p resolution with a frame rate of 30 FPS. The model demonstrates superior performance in handling complex, sequential character actions and realistic human interactions. While primarily available as a closed-source service via web and API, it is utilized across various industries, including film production, advertising, and digital storytelling.