Kling AI releases Kling 4.0 with 30 second video and stereo audio
A side by side test with Wan 3.0 and Seedance 2.5 shows that Kling 4.0 is equal to them and not above them.
Kling AI released Kling 4.0, an AI video tool. A user makes a maximum of 30 seconds of video in one step. The tool reads an input of 8,000 tokens. Kling AI writes that it also gets stereo audio and lip sync in more than 10 languages. A Flash version is for fast video at a low price. Kling AI gave no test results.
What is new
A user can give a maximum of 10 images to set the changes between parts of the video. Omni Reference uses a maximum of 15 input files in one step. Creative Recreation uses a video as a template and changes the item or the person in it.
A Flash version is for fast video at a low price, but only to 720p and 20 seconds. Kling 4.0 also works with MCP and CLI, which let AI agents control the tool.
The proof
In a side by side test with Wan 3.0 and Seedance 2.5, Kling 4.0 is equal and not above them. This applies to the sequence of shots, physics and voice.
What is not known
The API price is not known. HDR output and video of 120 seconds are not part of this release. The test can show a slow rate of change in AI video. The next Wan and Seedance releases will show more.
Insight: Agent access sets video tools apart
Kling 4.0 is not above the other tools in the test, but it adds MCP and CLI access for AI agents. ElevenLabs released a tool that changes voice agents. In some approval modes of an external AI tool, a change can go live with no check. It reads as a sign that, when quality is the same, teams keep the tool that agents can control. Set who gives the approval for agent changes.
Sources
Posted
