MiniMax H3 Max uses a default landscape canvas for text-only generation. For vertical short-form video, guide the canvas with a frame image or convert the output in post-production. The example below starts with a 480P landscape clip and converts it to 9:16 with ffmpeg.
Generate with this prompt → Select 6 seconds and 480P for this example; text-only output uses the default landscape canvas.
What We Actually Generated
We sent this exact prompt to minimax-h3-max in text-to-video mode:
A florist's hands wrap a bouquet of ranunculus and eucalyptus in brown kraft
paper on a sunlit marble counter, then tie it with twine in a neat bow; soft
morning window light, shallow depth of field, close-up vertical framing,
warm cozy mood, gentle paper-rustling motion
Generate with this prompt → Select 6 seconds and 480P.
Request body:
{
"model": "minimax-h3-max",
"input": {
"prompt": "A florist's hands wrap a bouquet of ranunculus and eucalyptus in brown kraft\npaper on a sunlit marble counter, then tie it with twine in a neat bow; soft\nmorning window light, shallow depth of field, close-up vertical framing,\nwarm cozy mood, gentle paper-rustling motion",
"duration": 6,
"resolution": "480P",
"aspect_ratio": "9:16"
}
}
We asked for 9:16. Here's the real output the API returned, byte-for-byte, no edits:
The historical 480P output measures 832×480. It uses the default landscape canvas despite the request's 9:16 value; it is separate from the newer 768P samples that measure 1344×768.
Text-Only Generation Uses a Default Landscape Canvas
HiAPI's model description states: "Text-only generation uses the default landscape canvas." The original request above follows that behavior: the source clip is landscape even though the prompt describes vertical framing and the request includes aspect_ratio: "9:16".
For a different canvas, supply first_frame_image or last_frame_image with the aspect ratio you need. Frame images and aspect_ratio are mutually exclusive, so omit aspect_ratio when using either frame field. You can also keep the landscape source and crop or pad it after generation, as shown below.
Turning It Vertical With One ffmpeg Command
Cropping the landscape frame to 9:16 would cut off half the bouquet. Instead we used a blurred-background pillarbox: scale a heavily blurred copy of the source to fill the vertical canvas, then overlay the original frame centered on top.
ffmpeg -y -i video.mp4 -filter_complex "\
[0:v]scale=720:1280:force_original_aspect_ratio=increase,crop=720:1280,gblur=sigma=20[bg];\
[0:v]scale=720:-1[fg];\
[bg][fg]overlay=(W-w)/2:(H-h)/2[out]" \
-map "[out]" -c:v libx264 -pix_fmt yuv420p -movflags +faststart video_vertical.mp4
One pass, no re-render, no lost frame content. Here's the result — same clip, now a real 720x1280 vertical file:
That's the version you'd actually publish to Reels, Shorts, or TikTok.
Pricing: What a Clip Like This Actually Costs
Minimax H3 Max bills per output second, tiered by resolution (confirmed on hiapi's pricing page):
- 480P — $0.072 / second
- 768P — $0.115 / second
Duration runs 5–15 seconds. A 6-second 480P request has a video budget of $0.432 at the current rate; 6 seconds at 768P is $0.69. Budget by generated seconds and confirm the final charge in your task billing records.
At the same duration, 768P costs about 1.60 times as much as 480P. Use 480P to test hooks and framing, then render selected shots at 768P; four or five successful 480P drafts still cost more than one 768P clip of the same length.
FAQ
Can Minimax H3 Max generate vertical video directly?
Text-only generation uses the default landscape canvas. For 9:16, use a frame image with that ratio and omit aspect_ratio, or convert the landscape output with a crop or blur-pad overlay.
How much does a 6-second clip cost? $0.432 at 480P ($0.072/sec x 6s), or $0.69 at 768P ($0.115/sec x 6s).
What's the difference between 480P and 768P beyond price? 768P is the higher-fidelity tier at roughly 1.6x the per-second cost of 480P — worth it for a final publish render, overkill for testing prompts and framing.
Does image-to-video avoid this orientation issue?
The API documentation supports frame images that determine the output canvas. Supply first_frame_image or last_frame_image and omit aspect_ratio. This article's videos show the text-only source and its ffmpeg conversion; they are not image-to-video samples.
Try It Yourself
The Minimax H3 Max model page has the live Playground if you want to test a prompt before writing any code, and our Minimax H3 Max API guide covers authentication and the full request/response shape end to end.
Generate with this prompt → Generate a 6-second 480P source, then use the ffmpeg conversion above for a vertical deliverable.
