Create videos with MiniMax H3 on PoloX using text, first and last frames, or image, video, and audio references. Choose 4–15 seconds and 768P or 2K output.
Capabilities
Video from a written brief
Describe the scene, action, and camera movement. MiniMax presents H3 as a multimodal generation system that can create video with native stereo sound.
First and last frame guidance
Guide a clip with a starting image, an ending image, or both. On PoloX, use Image to Video and describe the action or transition between frames.
Image, video, and audio references
H3 accepts mixed references. MiniMax demonstrates combining a character image, camera movement from a video, and vocals from audio. Describe which aspect of each reference should guide the new clip.
Sound alongside the picture
H3 jointly generates video and audio. Include dialogue, sound effects, or music directions in your brief, then review how the sound and action fit together in the result.
What you can do on PoloX
Text to Video
Write a prompt, choose an aspect ratio, set a duration from 4 to 15 seconds, and select 768P or 2K. This task does not require reference uploads.
Image to Video
Upload a first frame, a last frame, or both, and describe the motion. Choose duration and resolution. This PoloX form does not have a separate aspect-ratio selector.
Reference to Video
Upload image or video references and optionally add audio. Explain their roles in your prompt, then choose duration, resolution, and an explicit or adaptive aspect ratio.
FAQ
What is MiniMax H3?
MiniMax H3 is a multimodal generation system that works with text, images, video, and audio. On PoloX, it is available through Text to Video, Image to Video, and Reference to Video.
Which durations and resolutions can I choose?
All three PoloX H3 tasks offer whole-second durations from 4 to 15 seconds and output at 768P or 2K. MiniMax explains that its full 2K workflow regenerates a lower-resolution result using the original context.
Can I provide only a last frame?
Yes. PoloX Image to Video accepts a first frame, a last frame, or both. At least one frame is required. Each image can be JPEG, PNG, or WEBP, up to 30MB.
Can I generate from an audio reference alone?
The PoloX Reference to Video task requires at least one image or video reference. Audio is optional and must be paired with visual input. Use MP3 or WAV files, up to three files and 15MB each. Follow the official guidance of 2–15 seconds per audio clip and no more than 15 seconds of audio in total.
How should I combine reference files?
PoloX provides separate upload fields for up to nine images, three videos, and three audio files. Follow MiniMax’s combined limit of 12 reference files across all types; the individual field limits should not be added together. State what each file contributes to the desired result.
How are PoloX credits calculated?
Review the credit estimate in the generator before submitting. H3 pricing on PoloX depends on duration and resolution, with an additional reference-image charge where applicable. The displayed estimate reflects the current selections.