AI MiniMax H3 Video

INPUT

Input Type

Generate from text or keyframes, or guide the result with ordered references.

Text / Image References
Image / Video / Audio

Quality

Max follows the prompt more closely and usually looks better. It does not go above 1080p.

Standard Max
Better aesthetics and prompt adherence

Prompt

Describe the scene, motion, camera, style, and how each reference should be used.

Starting Image

Optional first frame. Leave blank to generate from text only.

0/1

Ending Image

Optional last frame. This is used only when a starting image is attached.

0/1

Reference Images

Upload up to 5 subject or style images. Order determines Image 1, Image 2, and so on.

0/5

Reference Videos

Upload up to 3 motion references, each 2-15 seconds. Video and audio references may total at most 15 seconds.

Max length is 15s. Longer files can be trimmed before running.

0/3

Reference Audio

Upload up to 3 audio references, each 2-15 seconds. Audio requires at least one reference image or video.

Max length is 15s. Longer files can be trimmed before running.

0/3

Resolution

Standard can render natively up to 4K. Max tops out at 1080p.

480p
Fastest 768p 1080p
Max only 2K
1440p 4K
2160p

Duration

Billed per second at the selected quality and resolution.

5 seconds 15 seconds

Aspect Ratio

Used for text and reference generation. Starting-image videos follow the uploaded image.

Auto

21:9

16:9

4:3

1:1

3:4

9:16

OUTPUT

Queued...

Your request will begin shortly...

Tip: You can safely leave this page and whenever the job is ready, come back and you will find the job in the "My Creations" section.

We did not forget about you. This step can sometimes take some time. But if you run a second request right after the first one completes, it is likely you will not have to wait as long since the model would be fully set up for use.

Running...

MiniMax H3 Video

INPUT

OUTPUT

Examples