Kling 4.0 stereo audio vs Sume's generate_audio flag

Kling 4.0 moves to two-channel stereo audio. Sume exposes generate_audio and per-model audio capability; check the catalog for what each model reports.

4 min readSume
All posts

Kling's 4.0 vs 3.0 post says audio moves from single-channel to two-channel stereo, with improved lip-sync accuracy. On Sume the audio control is the generate_audio request field, which defaults to the model's audio capability; the docs name stereo output for MiniMax H3 Max, not for Kling.

What does Kling say changed?

The post's comparison lists Kling 3.0 audio as single-channel and Kling 4.0 as two-channel stereo. It also says 4.0 adds languages such as Portuguese, German, French and Hindi to the 3.0 set of Chinese, English, Japanese, Korean and Spanish.

What audio options does Sume document?

Audio fields and claims in the Sume video docs, read 2026-10-01.
ItemDocumented behavior
generate_audioWhether to generate audio alongside the video; defaults to the model's audio capability
minimax-h3-max5-15 seconds with native stereo audio
gemini-omni-flash-1.1Native synced audio, 3-10 seconds
Audio referencesHonored by Seedance 2.x, Wan 3.0, MiniMax H3 and MiniMax H3 Max

Can I request Kling 4.0 stereo through Sume?

This post cannot confirm that. Check the live catalog response for a Kling 4.0 entry and its audio capability before you plan around it; see Kling 4.0 API availability.

How do I request audio explicitly?

Set generate_audio in the video request body for models that support it. See the video docs. For a model with stereo output, see MiniMax H3 stereo audio.

Sources

Related posts

More in Models

All Models posts

Written by Sume