Kling 4.0: 30-Second Clips, 10 Keyframes and Omni Reference

Kuaishou's Kling AI team announced Kling 4.0, a text-to-video model that doubles clip length to 30 seconds in a single pass, supports up to 10 keyframes per generation, and introduces Omni Reference with up to 15 multimodal references. It adds stereo audio with improved lip-sync and broader language support. Kling 4.0 Flash is in limited early access (up to 20 seconds, 720p), with the full model due in October 2026 and 4K HDR listed as coming soon.

• Maximum clip length doubled to 30 seconds in a single pass
• Up to 10 keyframes per generation for shot-by-shot control
• Omni Reference: up to 15 references from a pool of 50 files (30 images, 10 videos, 10 audio)
• Kling 4.0 Flash in early access at up to 20 seconds and 720p; full model due October 2026