Create 4-15 second 2K videos with native stereo audio from a prompt, first and last frames, or image, video, and audio references.
MiniMax H3 starts at 500 vid2vid credits for a 5-second 2K generation. Switch between Text, Image, and Video to use the matching H3 endpoint.
Open the generator when you are ready to upload a reference or write a prompt.
Open generatorWatch a finished 2K-style fantasy scene together with its creative direction and current vid2vid cost estimate.
A blue-lit warrior draws power from a luminous sword while fog, fabric, and subtle camera movement build a cinematic reveal.
Create a cinematic fantasy scene in cool blue light. A warrior kneels and grips a luminous sword as energy builds through the blade. Add drifting fog, natural fabric movement, dramatic contrast, and a slow controlled camera push-in with synchronized atmospheric sound.
H3 combines high-resolution output with directable motion, synchronized sound, and several ways to provide visual context.
H3 produces 2K video without a separate resolution selector, making it suitable for polished campaign and cinematic work.
Generate dialogue, ambience, effects, and music together with the visual result instead of treating sound as a separate pass.
Guide an image-to-video shot with a starting frame and an optional ending frame when the transition matters.
Use reference images, clips, and audio to control characters, motion, styling, wardrobe, backgrounds, or product presentation.
The model picker and API route use separate official KIE identifiers for each H3 workflow.
Text uses minimax-h3/text-to-video, frames use minimax-h3/image-to-video, and clips use minimax-h3/reference-to-video.
The duration control exposes every supported value from 4 through 15 seconds, while output remains fixed at 2K.
A 5-second text or image generation currently starts at 500 vid2vid credits; reference-video billing also includes source duration.
Choose H3 when the result needs more direction than a short prompt alone can provide.
Preserve source motion while changing a character, environment, material, lighting direction, or complete visual style.
Build polished product reveals, campaign variations, or social ads with controlled camera movement and synchronized sound.
Direct dialogue, character action, cinematic camera work, and atmosphere for short stories or animated sequences.
The selected H3 endpoint changes automatically when you switch the input mode.
Describe the subject, action, camera, lighting, dialogue, sound, and aspect ratio to create a new 4-15 second scene.
Upload a first frame and optionally a last frame, then describe the motion and transition between them.
Upload source footage and describe what to preserve and what to replace. H3 keeps the reference workflow in one request.
Answers about the Hailuo 03 model, inputs, output, duration, and billing.
Choose Text, Image, or Video, add a clear creative brief, and review the exact credit cost before generation.
