$20 of image-to-video: how many 5-second clips on Sume, by model
Twenty dollars buys 320 five-second Grok clips, 64 on Wan or H3 at 480p, 32 on Omni at 720p, or 19 on Kling 3 with audio. Sume prices, list x 1.25.
The short answer
Twenty dollars of Sume credit buys 320 five-second image-to-video clips on Grok Imagine Video 1.5, 64 on Wan 3.0 or MiniMax H3 at 480p, 53 on H3 at 768p, 32 on Gemini Omni Flash at 720p, and 19 on Kling 3 with audio. Counts are whole clips: $20 divided by the 5-second price, rounded down.
Clips per $20
The table does the division for you. Rates are the billed per-second rates, which are provider list times 1.25, and a 5-second clip is five times the rate.
| Model | Setting | 5 s price | Clips for $20 |
|---|---|---|---|
| grok-imagine-video-1.5 | 480p or 720p | $0.0625 | 320 |
| gemini-omni-flash-1.1 | 360p | $0.1875 | 106 |
| minimax-h3 | 480p | $0.3125 | 64 |
| wan-3.0 | 480p | $0.3125 | 64 |
| minimax-h3 | 768p | $0.375 | 53 |
| gemini-omni-flash-1.1 | 720p | $0.625 | 32 |
| kling-3 | silent | $0.70 | 28 |
| kling-3 | with audio | $1.05 | 19 |
Read the count as a ceiling
Counts are an upper bound for one specific plan: every clip is 5 seconds and nothing is re-run. In practice, plan for a share of attempts you throw away. If you keep one clip in three, divide the count by three.
Sound changes the comparison
Two columns of this table are not like the others. Grok and Kling silent return no sound. Omni and H3 always generate audio, so their price includes it. If your pipeline adds a soundtrack, the silent rows are the fair comparison for the $0.0625 and $0.70 prices.
A split that works
A sensible spend on $20 is a mixed batch: draft each shot on a cheap row, then finish the keepers on the model that fits the shot.
- $5.00 on Grok drafts: 80 clips at $0.0625
- $5.00 on Wan 480p tests of end-frame shots: 16 clips at $0.3125
- $10.00 on finals: 9 Kling 3 silent clips at $0.70 is $6.30, plus 5 Omni 720p clips at $0.625 is $3.125, for $9.425
One request shape
Each request is the same /v1/videos shape with a different model, a duration of 5 and your picture as a first_frame. Compare per-model options in the Videos API docs.
Sources
Related posts
More in Pricing
- A 20-slide narrated deck as video: about 71 cents on Sume
20 slides at 15 seconds each: 4,500 characters of narration is $0.21375 on Sume TTS and a 5-minute timeline is $0.50. Assumptions and caption limit.
- 2,000 product-name audio clips: 300,000 characters, MAI vs Sume
A catalog of 2,000 spoken product lines at 150 characters costs $4.50 on MAI Flash, $6.60 on MAI-Voice-2.1 and $14.25 on Sume, which cuts one file per line.
- A $25 video model bake-off: one 5-second prompt on six Sume tiers
How many 5-second test clips does $25 buy? From 35 up to 133 runs per tier, per-second Sume rates for Omni Flash, Wan, H3 and Kling 3 Pro.
- 25 six-second product loops: what 480p, 720p and 1080p cost on Sume
A catalog of 25 six-second loops costs $9.375 at Wan 480p and $37.50 at 1080p on Sume. Per-second rates and resolution steps, with fal list prices.
Written by Sume