Kling 3 Pro (Audio)

Overview

Created by Kuaishou, this model extends the visual output of Kling 3 Pro by generating synchronized audio tracks alongside video. It accepts text prompts and reference images, returning video and audio outputs in a single request. Kyma serves this model through an OpenAI-compatible endpoint with automatic failover if a serving path degrades. Prompt caching is supported, billing repeated prompt prefixes at this model’s cached input rate. New accounts receive a $0.50 free credit to test the endpoint. Every response includes exact billing in the usage.cost field and identifies the executed model via the X-Kyma-Model header. The model operates on a premium cost tier and generates output at a slower speed. It does not support structured outputs, reasoning chains, or vision analysis tasks. Input is limited to text and images, while output consists exclusively of video and synchronized audio tracks.

Specs

Field	Value
Model ID	`kling-3-pro-audio`
Best for	Cinematic clips needing diegetic sound, talking-head shots, ambient atmosphere
Context window	0
Input modalities	Text, Image
Output modalities	Video, Audio
Tool calling	Yes
Structured outputs	Yes
Prompt caching	Yes
Speed	Slow
Cost band	Premium
Release stage	Stable

Pricing

	Per 1M tokens
Input	$0.20
Output	$0.60

Use this when

Cinematic clips with ambient sound — Generate short film scenes that include synchronized environmental audio and background noise.
Talking head video generation — Produce character or spokesperson footage with matching dialogue audio tracks.
Product showcase with sound effects — Create promotional clips where visual actions align with generated sound effects.
Animated scene prototyping — Draft early video concepts with placeholder audio to evaluate pacing and mood.

Not ideal for

Do not use this model when you only need silent video generation or require fast, low-cost batch processing.

Pick something else when

You need silent video at a lower cost: use kling-3-pro.
You require faster generation speeds for drafts: use seedance-2-fast or veo-3-fast.
You want standard video without audio overhead: use kling-2.5-pro or hailuo-02-1080p.

Example

from openai import OpenAI

client = OpenAI(base_url="https://kymaapi.com/v1", api_key="ky-...")

response = client.chat.completions.create(
    model="kling-3-pro-audio",
    messages=[{"role": "user", "content": "..."}],
)

Agent query example

Ask the API which models fit, instead of hardcoding an id:

curl "https://kymaapi.com/v1/models?recommended_for=generate-video&tools=true&quality_tier=frontier-open"

FAQ

How does prompt caching work with this model? Kyma caches repeated prompt prefixes and bills them at this model’s cached input rate, reducing costs for iterative generation workflows. Can I use image inputs to guide the video generation? Yes, the model accepts both text prompts and reference images to control visual composition and style. What happens if the generation endpoint fails? Kyma automatically reroutes the request to an alternative serving path, ensuring the request completes without manual retry logic.

​Overview

​Specs

​Pricing

​Use this when

​Not ideal for

​Pick something else when

​Example

​Agent query example

​FAQ