OimiAI

MiniMax H3 AI Video Generator

MiniMax H3 is a multimodal AI video generator for creating 2K videos from text, first and last frames, or combined image, video, and audio references.

AI VideoMiniMax H3

FAQ

What is MiniMax H3 AI Video Generator?

MiniMax H3 is a multimodal AI video model that generates video from text prompts and can also use image, video, and audio references. It supports text-to-video, first-frame or first-and-last-frame video, and reference-based video generation.

What can I use as a reference for MiniMax H3?

In Oimi Canvas, MiniMax H3 supports up to 5 reference images, 3 reference videos, and 3 reference audio files. Use images for character, product, or style consistency; videos for motion and camera direction; and audio for voice or sound guidance.

Can MiniMax H3 generate video from a first and last frame?

Yes. Connect one image to create a first-frame video, or connect two images to control both the opening and ending frame. First-and-last-frame mode is separate from multimodal reference mode.

What video length and aspect ratios does MiniMax H3 support?

MiniMax H3 generates 2K videos from 4 to 15 seconds. In Oimi, choose from 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16.

Can I use audio by itself as a MiniMax H3 reference?

No. Audio reference must be paired with at least one reference image or reference video.

Create with more freedom

Explore the full workflow, prompts, and models on the AI infinite canvas, and let Agent help bring more of your ideas to life.

Try it now

Explore more AI workflows you can open and reuse in Canvas.

AI Chinese Aesthetic Wonderland: MiniMax H3 x MidJourney x Suno

AI Chinese Aesthetic Wonderland: MiniMax H3 x MidJourney x Suno

Using the latest MiniMax H3, MidJourney v8.2, and Suno, covering the entire process of images, videos, and soundtracks, it helps you quickly generate content with an Eastern aesthetic and a unified visual style. Whether it's a fairyland, ancient style, landscape, or atmospheric visuals, it can be efficiently implemented.

View template
MiniMax H3 Image To  Ghibli Art Video Generator

MiniMax H3 Image To Ghibli Art Video Generator

Use the latest MiniMax H3 video model to generate a Ghibli-style video from an anime-style image. First, upload an image or generate one in the Ghibli anime style using Midjourney, then convert that image into a video.

View template
Goodbye GTA 5 AI Video Generator – The Sunset Pier Farewell

Goodbye GTA 5 AI Video Generator – The Sunset Pier Farewell

Upload one photo — Michael, Franklin and Trevor hand you the keys on the sunset pier as you drive off into the GTA 6 era. No watermark, ready for TikTok&Reel.

View template
GTA 6 Loading Screen AI Video Generator – Your Photo Wipes In

GTA 6 Loading Screen AI Video Generator – Your Photo Wipes In

Upload one photo and AI turns it into your GTA 6 loading screen — the wipe transition, cel-shaded poster art, and trending-style sound in one 8-second clip. No watermark, ready for TikTok & Reel.

View template
Wassup in a Minute AI Generator – Live-Action Trio Dance

Wassup in a Minute AI Generator – Live-Action Trio Dance

Make the live-action Wassup in a Minute trend with you and up to two friends, from 1 to 3 photos — music included, 16:9, no watermark.

View template
BigXthaPlug Give It All AI Rap Video Generator

BigXthaPlug Give It All AI Rap Video Generator

Upload your photo or pet photo, AI casts you in the viral BigXthaPlug "Give It All" rap scene — the Big Sherman-style performance meme. No prompt, no watermark.

View template