Start with an image
Add a source image and describe its motion. You can also supply an end frame to guide where the clip finishes.
Animate an image or describe a scene with native audio, at 480p or 768p.

480p / 768p
Resolution options
3–15 s
Duration options
Text / image
Starting inputs
Inspect your own result for subject changes, camera movement, the final frame and audio. This page does not present generated H3 samples.
Add a source image and describe its motion. You can also supply an end frame to guide where the clip finishes.
Text-to-video offers seven aspect ratios, including 16:9, 9:16 and 1:1, alongside duration and resolution.
The existing H3 entry includes native audio. Describe the sound you want and listen to the generated result before keeping it.
One model choice with text and image starting points.
Open Image-to-Video when you have a reference to animate, or Text-to-Video for a scene described from scratch. Both H3 entries offer whole-second durations from 3 to 15 seconds and 480p or 768p output.
Image-to-video has no separate aspect-ratio selector for H3. Prepare the source composition for your intended frame; use the text entry when you want to choose an aspect ratio directly.
These are the options currently exposed by VideoAny. The names alone do not establish a quality or speed ranking.
| Entry | Current options | When to open it |
|---|---|---|
| MiniMax H3 | 480p / 768p, 3–15 seconds; text or image input, with an optional end frame for image-to-video. | Choose this entry for duration and resolution controls in the video studios. |
| MiniMax H3 Max | The separate AI video workspace uses H3 Max for credit-based text or image generation at a fixed 5 seconds and 768p. | Read the comparison first: successful paid videos and prompts in that workspace are public under your profile name. |
The main button selects the existing MiniMax H3 model.
Use a clear composition and add an optional end frame when the ending matters.
Specify the subject action, camera movement and the sound for one scene.
Choose 3–15 seconds and 480p or 768p. Check the displayed credit quote before generating.
Watch the full clip and listen to its audio. Revise one instruction at a time.
These are briefing examples, not tested output guarantees.
Example: The cyclist rolls slowly past the cafe, coat moving gently in the breeze.
Example: Hold a medium side view, then track smoothly alongside the cyclist.
Example: Soft tire noise on the road and quiet street ambience, with no dialogue.
For image-to-video, explain what should move and what should remain visually consistent.
Use an image for an existing subject or composition, and text for a new scene. Compare H3 and H3 Max before switching to their different workflows.
Continue with the supported H3 entry or explore a related model.
Animate a source image with the optional end-frame input.
Open H3 image-to-videoDescribe a scene and choose its aspect ratio.
Open H3 text-to-videoCompare the controls and the separate AI video workspace.
Read the comparisonRead the guide to the 4K option in the separate Seedance 2.5 entry.
Read the 4K guideExplore another video model page, available in English.
View Wan 3.0 in EnglishExplore its separate model page, available in English.
View Seedance 2.5 in EnglishBrowse the model pages available on VideoAny.
Browse modelsYes. Open Text-to-Video with MiniMax H3 and describe your scene.
The H3 studio entries offer 3–15 seconds in whole-second steps, at 480p or 768p.
Yes, the H3 image-to-video entry accepts an optional end frame in addition to the source image.
H3 Max is used in the separate AI video workspace. Its current credit-based flow uses 5 seconds and 768p; see the comparison for the workflow and publication details.
Check your inputs and review the finished clip before sharing.
Start from an approved image, describe one scene, and choose the available duration and resolution.
Open MiniMax H3