MiniMax H3 reference images: the sixth one costs 10 cents extra
MiniMax H3 reference-to-video includes five images; Sume adds $0.10 per image after that ($0.08 list x 1.25). A 10 s clip with 8 images at 768p is $1.05.

MiniMax H3 and H3 Max take reference images for reference-to-video, which is how you keep a character or product consistent across a clip. Sume's pricing code sets the cost rule: reference-to-video adds $0.08 per image after the fifth, at provider list. With Sume's 1.25 multiple, each extra image is $0.10.
The math for a 10-second 768p clip
MiniMax H3 lists $0.06 per second at 768p in the Sume pricing code. Ten seconds is $0.60 list, or $0.75 billable. Eight reference images means three beyond the free five: 3 x $0.08 = $0.24 list. Total list $0.84, billable about $1.05.
| Reference images | Extra images | List | Sume billable |
|---|---|---|---|
| 1 to 5 | 0 | $0.60 | $0.75 |
| 6 | 1 | $0.68 | $0.85 |
| 8 | 3 | $0.84 | $1.05 |
| 10 | 5 | $1.00 | $1.25 |
Choose images by value
The fal page for H3 Max shows the per-second tiers and a 40 percent promotion that ends October 15 (read 2026-10-05); Sume prices at list, not at the promotion.
- Five is the free ceiling. Spend the first five on the faces, product and background that must not drift.
- A sixth image costs about the same as one extra second of H3 Max at 768p ($0.10).
- Check the live catalog for the reference limits of the exact model id you call.
Call
{
"model": "minimax-h3",
"prompt": "The two characters toast across the table",
"duration": 10,
"resolution": "768p",
"reference_image_urls": [
"https://example.com/a.png",
"https://example.com/b.png"
]
}
Sources
Related posts
More in Pricing
- A monthly AI video budget calculator in Python using Sume rates
Plan a month of 480p drafts, 720p posts and audio clips: $97.70 on Sume. A short Python calculator with per-second rates and captions you can edit.
- Monthly budget for 25 sales reps sending personalized avatar videos
25 reps x 20 clips a month at 20 seconds each, priced on Sume standard, plus and max, with one shared avatar and what changes if each rep has their own.
- Monthly budget for 30-second Wan 3.0 clips: 4, 12 or 30 a month
Four 30 second Wan 3.0 clips a month cost $15 at 720p on Sume, twelve cost $45 and thirty cost $112.50. Tables for 480p and 1080p, plus a draft-and-final plan.
- Museum audio guide: 80 stops in three languages, TTS cost
An 80-stop audio guide at 700 characters per stop in three languages is 168,000 characters: $2.52 on MAI Flash, $3.70 on MAI-Voice-2.1, $7.98 on Sume.
Written by Sume