Kling 4.0 leak: 30s video generation, 15 reference elements, 1080p

xiaohu · x · 2026-09-24

Kling, quiet for months, is reportedly set to return with Kling 4.0. The leak claims up to 30-second video generation (extendable to 60s), support for up to 15 reference elements, 1080p resolution, multilingual audio-visual sync with up to 3 voice references, and possibly a long-video mode stretching to 120 seconds at 1080p. Unconfirmed rumor for now.

Related event: Kuaishou Teases Kling 4 as Leaks Point to 120-Second Video and 4K Images(5 posts)→

Original post →

More from Multimodal

Multimodal channel →