Executive Summary
Kling's suite of AI video generation models, including the new state-of-the-art Kling 3.0, are now available in beta through the AI Gateway and AI SDK V6. The models enable users to generate high-quality, cinematic videos from text prompts, still images, or specified start-and-end frames. Key capabilities of the new models include multishot narrative generation from a single prompt, synchronized audio creation, and 1080p resolution output for paid users on Pro and Enterprise plans.
Key Takeaways
* Product & Availability: Kling video models, including versions 3.0, 2.6, and 2.5 turbo, are now accessible via the AI Gateway and programmatically through AI SDK V6. The service is currently in beta for Pro, Enterprise, and paid AI Gateway users.
* Core Functionality: The models generate video from text (Text-to-Video) and animate still images (Image-to-Video). They are noted for realistic motion, physics, and high-resolution output up to 1080p in "pro mode."
* Kling 3.0 Key Features: The latest version introduces advanced capabilities, including:
* Multishot Narratives: Generates a complete, multi-scene video from a single narrative prompt, with the AI intelligently cutting between shots.
* Audio Generation: Creates synchronized sound effects and ambient audio to accompany the video.
* First & Last Frame Control: Allows users to specify start and end images for precise scene transitions, ideal for effects like time-lapses.
* Access Methods: Users can interact with the models in two ways: programmatically using `experimental_generateVideo` in the AI SDK, or through a no-code Playground environment within the AI Gateway.
Strategic Importance
This integration makes cutting-edge, cinematic AI video generation accessible to developers and businesses through a unified API, positioning AI Gateway as a competitive hub for state-of-the-art multimodal models.