Reference images
Up to 9 images set the subject and style. You name them in the prompt as Image 1, Image 2 and so on.
Bring a picture to life
MiniMax H3 is a frontier video model. This endpoint generates 2K video from multimodal references up to 9 images for subject and style, 3 video clips for motion, and 3 audio clips each cited in the prompt by order, keeping subjects consistent while following the referenced motion and audio.
Opens the video studio with MiniMax H3 Reference to Video ready to use. You can look around as a guest. Running it takes the Pro plan or higher.
What it makes
Bring a picture to life
Animates a picture into a short video.
What you need
Optional: the picture to animate, a video to start from and an audio clip.
What you get
A short video that starts from your picture
MiniMax H3 Reference to Video makes a video from reference files. You can supply up to 9 images for subject and style. In Reference to Video mode you can also add up to 3 video clips for motion and up to 3 audio clips. Your prompt points to each file by its order, such as Image 1, Video 1 or Audio 1.
This model has two modes. In the mode called Ref to Video (Ref is short for Reference), you must add reference images and a prompt, and Duration runs from 8 to 10 seconds. In the mode called Reference to Video, only a prompt is required. Reference images, videos and audio are options there, along with Aspect Ratio and Resolution, and Duration runs from 8 to 15 seconds. Images, videos and audio clips together can be at most 12 files. Each video or audio clip must be 2 to 15 seconds long. All the video clips together can run at most 15 seconds, and the same goes for the audio clips. Audio cannot be the only reference, so add at least one image or video with it.
The price is 81 credits. MiniMax H3 works from a prompt with an optional Image and End Image, and costs about 2500 credits for 5s.
Same job: bring a picture to life. Compare before you choose.
AI video generation model that creates smooth, coherent videos from a first frame and last frame, generating the motion and transitions between them
Turbo is the model to use when you feel the need for speed. Turn your image to stunning video up to 3x faster – all with high quality outputs.
Turn photos into mind-blowing, dynamic videos. Your images can can come to life with sharp details, impressive character control and cinematic camera moves.
Expressive facial performance, natural speech-expression coordination, realistic body motion, and accurate audio-video synchronization with DaVinci-MagiHuman model
Up to 9 images set the subject and style. You name them in the prompt as Image 1, Image 2 and so on.
In Reference to Video mode, video clips of 2 to 15 seconds each guide the motion, with at most 15 seconds in total.
In Reference to Video mode, audio clips of 2 to 15 seconds each can go along with an image or a video.
Images, videos and audio clips together can be at most 12 files.
Upload reference images for the subject and style. In Reference to Video mode you can add video and audio clips too.
Refer to your files by their order, such as Image 1, Video 1 or Audio 1.
Choose a Duration, and in Reference to Video mode also Aspect Ratio and Resolution, then generate.
Add images of your subject as references and name them in the prompt.
In Reference to Video mode, add a reference video and name it in the prompt to guide the motion.
In Reference to Video mode, add an audio clip together with an image or a video, and name it in the prompt.
Quick answers to help you get the most out of MiniMax H3 Reference to Video in FunAI Studio.