Kling 3.0: Native 4K, Multi-Shot Video, and Built-In Audio

Kuaishou released Kling 3.0 on February 4, merging the capabilities of Kling 2.6 and Kling O1 into a single unified system. The model is available on Kling's platform for Ultra subscribers and exclusively via API through fal.ai.

• Native multi-shot storyboarding generates up to 6 distinct shots in one clip. Each shot gets its own prompt and duration, with total clips running up to 15 seconds. Characters maintain spatial continuity across different camera angles, a step beyond previous single-clip generation.

• Native 4K resolution and integrated audio generation. Video generates natively at up to 3840x2160 (not upscaled) at 60fps. Audio, including dialogue, ambient sound, and voice tone control, generates simultaneously with the video rather than in a separate pass. Multilingual support covers English, Chinese, Japanese, Korean, and Spanish with multi-speaker dialogue.

• Reference-to-video enables editing capabilities. Creators can change backgrounds, modify clothing, insert or remove people, and reshape scenes while preserving the original structure and character identity.

• API pricing through fal runs $0.17-$0.39 per second depending on tier and features. Standard without audio costs $0.168/second; Pro with audio and voice control costs $0.392/second. A 5-second Pro clip with full audio runs roughly $1.96, about 2-3x more expensive than Kling 2.6.