Hailuo 02 (768p) - Kyma API

Overview

Created by Minimax, this model generates video from text prompts or reference images at 768p resolution. It operates in a strong quality tier with a balanced cost structure, making it suitable for consistent content pipelines. On Kyma, it runs through an OpenAI-compatible endpoint using a single API key. The platform handles automatic failover if a serving path degrades, and every response returns the exact cost in usage.cost along with the routed model in the X-Kyma-Model header. Prompt caching is supported, billing repeated prefixes at this model’s cached input rate. The model operates at a slower speed tier and does not support reasoning, vision, or structured outputs. It is strictly a video generation endpoint, so it will not process conversational tasks or return JSON schemas.

Specs

Field	Value
Model ID	`hailuo-02-768p`
Best for	Brand b-roll, product motion, character shots at production quality
Context window	0
Input modalities	Text, Image
Output modalities	Video
Tool calling	Yes
Structured outputs	Yes
Prompt caching	Yes
Speed	Slow
Cost band	Balanced
Release stage	Stable

Pricing

	Per 1M tokens
Input	$0.20
Output	$0.60

Use this when

Brand B-Roll Generation — Generate consistent background footage for marketing campaigns and social media posts.
Product Motion Demos — Turn static product images into short, high-quality motion clips for e-commerce.
Character Animation Shots — Produce cinematic character movements from text descriptions or reference frames.

Not ideal for

Do not use this model for real-time video generation, interactive applications, or tasks requiring structured JSON outputs or reasoning.

Pick something else when

You need faster generation speeds: use veo-3-fast.
You need higher resolution output: use hailuo-02-1080p.
You require audio generation alongside video: use kling-3-pro-audio.
You need maximum visual fidelity regardless of cost: use kling-3-pro.

Example

from openai import OpenAI

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

response = client.chat.completions.create(
    model="hailuo-02-768p",
    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=strong"

FAQ

Does this model support image-to-video generation? Yes, you can pass an image URL in the image_url field to generate video from a reference frame. How does prompt caching work with this endpoint? If you reuse prompt prefixes across requests, Kyma caches them and bills that cached portion at this model’s cached input rate. Will I get charged if a request fails? No, Kyma routes requests through automatic failover and only bills successful generations, with exact costs returned in the usage.cost field.

​Overview

​Specs

​Pricing

​Use this when

​Not ideal for

​Pick something else when

​Example

​Agent query example

​FAQ