MiniMax H3 4K hero clip vs 768p cutdowns: holiday ad cost
A 10-second MiniMax H3 clip is $0.75 at 768p, $1.625 at 2K and $2.00 at 4K on Sume. Five free reference images are included; extra ones are $0.10 each.

On Sume, MiniMax H3 costs $0.075 per second at 768p, $0.1625 at 2K and $0.20 at 4K, so a 10-second hero clip is $0.75, $1.625 or $2.00. H3 takes 5 to 15 seconds and the first 5 reference images are free, then $0.10 each.
Rates
Per-second figures are the provider list rate times the 1.25 Sume margin that the docs describe, as of 2026-10-08. Read capabilities from GET /v1/video-router/models before you bill a batch, because rates and envelopes can change.
| Tier | Per second | 10 s | 15 s |
|---|---|---|---|
| 480p | $0.0625 | $0.625 | $0.9375 |
| 768p | $0.075 | $0.75 | $1.125 |
| 2K | $0.1625 | $1.625 | $2.4375 |
| 4K | $0.20 | $2.00 | $3.00 |
A hero plus cutdown plan
Make one 4K hero at 10 s ($2.00) and ten 768p cutdowns at 5 s each (10 x 5 x 0.075 = $3.75), for $5.75 in total. All five of the main ratios (16:9, 9:16, 4:3, 3:4, 1:1) plus 21:9 are listed for H3.
- H3 Max is a separate model: 480p $0.0625, 768p $0.10, 1080p $0.20.
- H3 Max 1080p is a latent refinement from native 768p.
- Reference images: 5 free, then $0.10 each.
Reference image cost
With 8 reference images on one H3 call, 3 are charged: 3 x $0.10 = $0.30 on top of the clip. Keep the reference set at 5 or fewer for bulk runs, which also keeps the estimate stable.
Choosing between H3 and H3 Max
H3 is a better fit for 4K and 2K deliverables, since H3 Max tops out at 1080p. H3 Max at 768p costs $0.10 per second against $0.075 for H3, and is the faster variant. For a pure 768p cutdown set, H3 is cheaper: ten 5-second clips are $3.75 on H3 and $5.00 on H3 Max.
H3 accepts 5 to 15 seconds, so shorter bumpers need another model.
- H3 10 x 5 s at 768p: $3.75.
- H3 Max 10 x 5 s at 768p: $5.00.
- Difference: $1.25.
Before you scale
Reference images are the hidden line item. A hero shot of a product from five angles is free on H3, and the sixth image onward is $0.10 each. If a catalog template adds a logo, a texture and three product angles, that is five images and free. Add two lifestyle references and you pay $0.20 per clip, or $20.00 across 100 SKUs.
Sources
Related posts
More in Models
- MiniMax H3 defaults to 2K at MiniMax; on Sume the default is 768p
MiniMax's H3 page says 2K is the default. Sume's minimax-h3 runs native 480p or 768p with a 768p default; 2K and 4K are billed upscales.
- MiniMax H3 on a 12 GB RTX 3060: 480p only, 42 GB, and the price
MiniMax says 12 GB of VRAM runs 480p H3 clips with audio from a pruned int8 checkpoint of about 42 GB. What that buys, and a 480p clip on Sume costs.
- Mistral Large 4 open weights: Oct 31 ETA and what to do now
Mistral's own Hugging Face page lists Mistral Large 4 as upcoming with an October 31 ETA. What is known, what is not, and why Sume's agent run is unaffected.
- Moroccan Arabic speech to text: open-source Voxtral model vs Sume STT
A Moroccan open-source stack pairs language ID with Voxtral ASR for dialect code-switching. What Sume STT offers for Arabic audio, and what it lacks.
Written by Sume