Music visualizer clip: MiniMax H3 Max, image plus audio reference
Make a 10-second visualizer from cover art plus a track excerpt with minimax-h3-max on Sume: $1.00 at 768p; audio alone is not a valid reference.

For a 10-second visualizer, send the cover art as an image_url and the track excerpt as an audio_url in input_references to minimax-h3-max. At 768p it costs $1.00 on Sume ($0.62 at 480p, $2.00 at 1080p). An audio reference cannot be the only reference, so the image is required; the model limits are nine images, three videos and three audio clips, at most twelve in total (Video Router, read 2026-10-05).
What is the request?
The shape comes from the Video generation guide: input_references entries are tagged image_url or audio_url. H3 Max always produces stereo audio, and there is no generate_audio toggle.
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"minimax-h3-max","prompt":"Abstract waveform ripples radiate from the album cover, neon edges pulse with the beat","duration":10,"resolution":"768p","aspect_ratio":"1:1","input_references":[{"type":"image_url","image_url":{"url":"https://example.com/cover.jpg"}},{"type":"audio_url","audio_url":{"url":"https://example.com/excerpt.mp3"}}]}'How does the price change with length and resolution?
H3 Max is billed per second by resolution: list $0.05, $0.08 and $0.16 times 1.25. Native renders are 480p and 768p; 1080p is a refinement from the 768p render.
| Length | 480p | 768p | 1080p |
|---|---|---|---|
| 5 seconds | $0.31 | $0.50 | $1.00 |
| 10 seconds | $0.62 | $1.00 | $2.00 |
| 15 seconds | $0.94 | $1.50 | $3.00 |
How do I poll and download the result?
A submit returns 202 with id, polling_url, status and model. Poll GET /v1/videos/{id} until status is completed (the other terminal values are failed and cancelled), then download from GET /v1/videos/{id}/content?index=0 with the same bearer key. A sound-only reference with no image comes back as a 400.
curl https://api.sume.com/v1/videos/JOB_ID -H "Authorization: Bearer $SUME_API_KEY"
curl -L "https://api.sume.com/v1/videos/JOB_ID/content?index=0" \
-H "Authorization: Bearer $SUME_API_KEY" --output clip.mp4Will the visuals follow the beat?
Do not promise a beat-accurate result: the audio reference conditions the clip, it is not a beat grid. Check the first render and describe the rhythm in the prompt. For the audio side, see directing the audio of MiniMax H3.
Sources
Related posts
More in Use cases
- Napkin sketch to product render: one reference on GPT Image 2.5
Turn a photographed napkin sketch into a product render with one input reference on GPT Image 2.5. Set aspect_ratio auto to keep the shape. Code and prices.
- New Year countdown video music: build to midnight with a timed prompt
Write a timed Lyria 3.5 prompt that builds for 50 seconds and hits at midnight. $0.125 a take on Sume, then a Timeline render at $0.10 under your countdown.
- Nine product angles as references: MiniMax H3 Max on Sume
How to send up to nine product photos as reference images to minimax-h3-max on Sume, what the vendor limit is, and where Sume's own limits live.
- One 90-second ad, three voice beats: three TTS jobs and a concat
Splitting a 1,350-character read into hook, proof and call-to-action jobs costs 8 cents plus 1 cent to join, against 7 for one job. When that is worth it.
Written by Sume