# Kling 3 Pro

> Flagship Kling. Photoreal humans, smooth motion, sharper than 2.5. T2V or I2V via image_url. For native audio, use kling-3-pro-audio.

Human version: https://kymaapi.com/models/kling-3-pro
Live JSON: `GET https://kymaapi.com/v1/models` (no auth required)

## Facts

- **Model ID**: `kling-3-pro` — pass this as `model` in the request body
- **Creator**: Kuaishou
- **Released**: 2026-02-05
- **Price**: $0.1512 / second

## Call it

```bash
curl https://kymaapi.com/v1/videos/generations \
  -H "Authorization: Bearer $KYMA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "kling-3-pro", "prompt": "..."}'
```

## Positioning

Kling 3 Pro is Kuaishou's flagship video generation model, optimized for photorealistic human motion and high-fidelity cinematic output. Reach for it when you need premium text-to-video or image-to-video generation and can accept slower inference speeds.

## About

This model generates video from text prompts or input images via an image_url parameter. It delivers sharper detail and smoother motion than its predecessor, with a focus on photorealistic human subjects and cinematic framing.

On Kyma, requests route through an OpenAI-compatible endpoint with automatic failover to maintain reliability. Prompt caching is supported, billing repeated prompt prefixes at 10% of the input rate. Every response returns the exact compute cost in usage.cost and identifies the active model in the X-Kyma-Model header.

The model operates on a premium cost tier and runs at a slower inference speed compared to fast-tier alternatives. It does not natively generate audio; for synchronized soundtracks, you must route to the dedicated audio variant.

## Use cases

- **Cinematic Character Shots** — Generate high-fidelity video clips focused on realistic human faces and smooth motion.
- **Brand Hero Videos** — Produce premium marketing assets from text prompts or reference images.
- **Image-to-Video Conversion** — Animate static reference frames into continuous, photorealistic video sequences.
- **Premium Visual Prototyping** — Test high-quality video concepts before committing to full production pipelines.

## Not ideal for

Do not use this model for real-time applications, low-budget batch generation, or workflows that require native audio synthesis.

## Pick something else when

- You need faster inference speeds → `seedance-2-fast`, `veo-3-fast`
- You require native audio generation → `kling-3-pro-audio`
- You are generating lower-resolution drafts → `hailuo-02-512p`, `hailuo-02-768p`
- You need a cheaper alternative for bulk testing → `kling-2.5-pro`

## See also

- All models: https://kymaapi.com/models.md
- Pricing: https://kymaapi.com/pricing.md
- Other models by Kuaishou: https://kymaapi.com/models?q=Kuaishou