Kuaishou's Kling 4.0 video model is now fully available: refreshed creation studio plus MCP and CLI access, with 4K 10-bit HDR and enterprise API coming

Why it matters

【Key points】

On Oct 15 (15:00 UTC) Kuaishou's Kling AI announced that its Kling 4.0 video-generation model is "now fully available." What changes from that day, per Kling: Kling 4.0 is usable through a refreshed creation interface on kling.ai, and is accessible via MCP and a CLI so it can be wired into agents and automated workflows. Listed as "coming soon": 10-bit HDR video output at 4K and 1080p, and API access for enterprise customers and developers.

Kling 4.0 Flash shipped a few days earlier while 4.0 was in beta. Partner platform Runware previously described the specs as up to 30-second outputs, up to 10 keyframes and 15 references in a single request, and 4K with stereo audio. These are the partner's claims; Kling's own post does not itemize them. About 14 hours after posting, the official announcement had ~82K views and ~850 likes.

【Verifiable】

Kling announcement: https://x.com/Kling_ai/status/2110747537187709424

Kling site: https://kling.ai

Runware spec preview: https://x.com/runware/status/2110414581449314544

【Our read】

1. A leading Chinese video-model vendor is pushing its new generation straight to all users and exposing it through MCP and a CLI at the same time, a sign that video generation is being treated as a tool agents can call, not just a web app for people.

2. If the 30-second length and multi-keyframe, multi-reference control hold up, the main problem it addresses is keeping characters and motion consistent across long shots, the biggest blocker for commercial short-form work. For now the quantified specs come mostly from partners and user tests, so treat it as "shipped, pending detailed testing" until the formal API and pricing land.

3. HDR and the enterprise API are not live yet. Professional film and advertising users can't really integrate it until both arrive.