OpenAI's latest video generation model with enhanced quality
Google's latest model with exceptional quality output
Fast and efficient model with flexible aspect ratio controls
When this option is enabled, the output video may be selected by AI Generate and published to the Explore.
The workbench preview is a site sample, not a verified output from this model.
MiniMax H3 supports 4–15 second video generation from text, first or last frames, and reference media. Both input and output video durations are billed; the first five reference images and input audio add no separate charge.
Start from a text prompt. Use image mode to add reference images.
Choose 768P / 2K and a duration of 4–15 seconds.
Up to 9 images, including uploaded references. Up to 3 reference videos. Up to 3 audio references.
Choose the aspect ratio, resolution and, for video, duration. Review the credit estimate before submitting.
Write a prompt for a new scene, or switch to image mode and add a reference. Choose the matching reference mode when using frames or other media.
Choose the aspect ratio, resolution and, for video, duration. Review the credit estimate before submitting.
Sign in and submit once. Follow the task in the preview or history, then download the finished result. Processing time depends on the provider queue.
Prepare a scene with several visual references or test a frame-led transition. Account for extra image charges when using more than five images.
Start with a close-up of a plant on a windowsill, then slowly pull back to reveal the room. Keep the plant and lighting consistent.Only add references that help define the subject, framing or motion. Remove unrelated files before generating.
Up to 9 images, including uploaded references. Up to 3 reference videos. Up to 3 audio references.
| Resolution | Credits / output second |
|---|---|
| 768P | 8 |
| 2K | 13 |
Credits = the resolution rate × (input video duration + output duration), plus 4 credits for each reference image after the first 5. Input audio is free.
Example at 768P: 5 seconds without reference video costs 40 credits.
The final total is rounded up to a whole credit. View plans
Start from a text prompt. Use image mode to add reference images. Up to 9 images, including uploaded references. Up to 3 reference videos. Up to 3 audio references.
Choose 768P / 2K and a duration of 4–15 seconds.
Credits = the resolution rate × (input video duration + output duration), plus 4 credits for each reference image after the first 5. Input audio is free. The final total is rounded up to a whole credit.
Your submitted task appears in generation history. You can return to check its status and open completed results.