Overview

MiniMax H3 Reference to Video API

MiniMax H3 reference to video takes your own material as the reference for a new clip — characters, products, motion and voice — and returns 5 to 15 seconds of video with native stereo audio. Up to 9 images, 3 videos and 3 audio clips can travel with a single request.

On the GoEnhance Open API it is the hailuo_h3_refs model on the same REST endpoint as the rest of the video catalogue. This endpoint runs on MiniMax's official channel, so its tiers are 768P and 2K.

Prefer a visual editor over writing code? Try MiniMax H3 Reference to Video on goenhance.ai.

Specifications
Modes
Reference to video
Reference inputs
Up to 9 images + 3 videos + 3 audio clips
Duration
5–15s
Resolution
768P · 2K
Aspect ratio
adaptive · 21:9 · 16:9 · 4:3 · 1:1 · 3:4 · 9:16
Audio
Native stereo audio
Prompt
Up to 7,000 characters
Endpoint
POST /api/v1/videos/generations
Price
$0.099 / sec
Frequently asked questions
What can I use as a reference?

Up to 9 images, 3 videos and 3 audio clips in one request, passed as ref_imgs, ref_videos and ref_audios. Use them to carry a character, a product, a camera move or a voice into the generated clip.

How long can the output be?

Any whole number of seconds from 5 to 15, the same as the other MiniMax H3 endpoint. Requests without a duration default to 5 seconds.

Can I combine references with a first and last frame?

No. MiniMax treats frame control and reference material as mutually exclusive, which is why GoEnhance exposes them as two endpoints: use hailuo_h3 for first/last frame, and this model for reference-driven generation.

What does it cost?

It is billed per second of output, by resolution — see the pricing table on this page. You are charged once per task at submission, and refunded in full if generation fails.

MiniMax H3 Reference to Video API | GoEnhance Open API