MiniMax H3 Studio
Build a complete scene from a prompt
H3 generates native stereo audio. Include dialogue, music, ambience, and timing directly in your prompt.
MiniMax H3 Video Generator
Create complete audiovisual scenes from a prompt, animate exact first and last frames, or combine image, video, and audio references for controlled character, motion, camera, and voice direction.
Choose how your video begins
MiniMax H3 understands each source as part of the same direction. Start simple, or assign different references to identity, motion, cinematography, and sound.
Text to Video
Write the scene, performance, camera plan, pacing, dialogue, and sound design. H3 builds the full clip without source media.
Image to Video
Animate one image as the exact opening frame, or provide both first and last frames for a more deliberate transition.
Reference to Video
Mix images, video, and audio to preserve identity, borrow movement, match a voice, or apply targeted creative changes.
Direct the complete shot, not just the first frame
Visual action, camera movement, typography, dialogue, music, and environmental sound can all be described together. Reference assets give the model concrete anchors when words alone are not enough.
Native stereo audio
Direct dialogue, music, ambience, footsteps, and sound effects in the same prompt as the visuals, so picture and sound develop together.
Stronger subject continuity
Use reference images to anchor a character, product, wardrobe, or visual identity across changing shots and camera movement.
Motion and camera references
Bring in reference video for choreography, pacing, shot language, and editing rhythm instead of describing every movement from scratch.
Production-ready control
Generate 4–15 second clips in 768P or 2K, with flexible aspect ratios and detailed instructions for timing, typography, and brand elements.
Pay for the video time H3 processes
Output duration and reference-video duration use the same selected resolution rate. Five reference images are included, and audio references do not add cost.
From source material to finished MP4
Choose a mode
Start from text, opening and closing frames, or a multimodal reference set.
Direct every element
Describe timing, camera, performance, continuity, dialogue, music, and restrictions.
Generate and download
Select duration and resolution, review the live credit estimate, then download the finished MP4.
MiniMax H3 questions, answered
Have a different question and can't find the answer you're looking for? Reach out to our support team by sending us an email and we'll get back to you as soon as we can.
What is MiniMax H3, and is it the same as Hailuo 03?
Yes. MiniMax H3 is also presented as Hailuo 03 or Hailuo 3.0. It is a multimodal video model that can understand text, images, video, and audio within one generation workflow.
Which generation mode should I choose?
Use Text to Video when you are starting from an idea, Image to Video when an opening frame or first-and-last-frame transition must be preserved, and Reference to Video when identity, motion, camera language, or voice should come from separate source files.
How much does MiniMax H3 cost?
768P costs 31 credits per generated or reference-video second, while 2K costs 50 credits per second. The first five reference images are free, each additional image costs 16 credits, and reference audio is free.
How are reference videos billed?
Reference-video duration is added to the generated-video duration at the selected resolution rate. The server reads the uploaded video metadata and calculates the final credit requirement before the task is submitted.
Can MiniMax H3 generate audio?
Yes. MiniMax H3 can generate native stereo audio with the video. Describe dialogue, music, ambience, and sound cues in the prompt. You can also upload audio references in Reference to Video mode at no additional credit cost.
What files can I use as references?
Reference to Video accepts up to nine JPG, PNG, or WebP images, three MP4 or MOV videos, and three MP3 or WAV audio files. Image to Video accepts a first frame and an optional last frame.
Build your next scene with MiniMax H3
Move from an idea to a complete audiovisual clip, or use real source material to guide identity, motion, camera behavior, and sound.
Open MiniMax H3 Studio