A 2-second 9:16 hook clip costs 25 cents, and only Wan 3.0 makes it
Wan 3.0 is the only Sume text-to-video row that starts at 2 s. At 720p a 2 s clip bills 25 cents; 3, 4 and 5 s bill 38, 50 and 63. Other rows start at 3 to 5 s.

Wan 3.0 is the only prompt-driven video row on Sume that accepts a 2-second clip. At 720p a 2-second clip bills 25 cents, and the whole short end of the range is cheap: 3 seconds is 38 cents, 4 seconds is 50 cents and 5 seconds is 63 cents. Every other text-to-video row starts at 3 seconds (Gemini Omni Flash 1.1), 4 seconds (Seedance and Kling 3) or 5 seconds (MiniMax H3 and H3 Max).
Wan 3.0 short-clip prices
Billable prices, which are list price times 1.25 rounded up to the cent.
| Length | 480p | 720p | 1080p |
|---|---|---|---|
| 2 s | $0.13 | $0.25 | $0.50 |
| 3 s | $0.19 | $0.38 | $0.75 |
| 4 s | $0.25 | $0.50 | $1.00 |
| 5 s | $0.32 | $0.63 | $1.25 |
Arithmetic
The price is linear in seconds. At 720p the rate is 12.5 cents a second, so 2 s is 25 cents, 3 s is 37.5 cents rounded up to 38, and 5 s is 62.5 cents rounded up to 63. At 480p the rate is 6.25 cents a second, so 2 s is 12.5 cents, rounded up to 13. The rounding is applied to each clip, so ten 3-second clips cost 10 x 38 = 380 cents, not 375.
When a 2-second clip is useful
A short clip is a building block. A hook of two seconds followed by a longer body can be assembled in Timeline 1.0, which renders at $0.10 per started output minute, so a hook and body of under a minute adds one $0.10 step. A 2-second Wan clip plus a 10-second Wan clip at 720p is 25 + 125 = 150 cents, and the join is another 10 cents, which makes $1.60.
Check the platform side before you plan around it. TikTok's non-Spark page gives no minimum duration in the facts used here, and no short-clip claim is made about how a 2-second clip performs. The 2-second floor is simply what the Sume catalog lists for Wan 3.0, so read supported_durations from GET /v1/videos/models before you build a form around it.
Check before you build on it
The catalog and the router docs are the source for the 2-second floor. If a future change adds a lower floor to another row, supported_durations in the live models list will show it. Build any form or script on that field, not on the number in this post.
Also note that the 2-second clip is a plain generation, so the same audio and aspect rules as other Wan 3.0 clips apply. It accepts 9:16, 16:9, 4:3, 3:4 and 1:1.
Hook test budget
Ten different 2-second hooks at 720p cost 10 x 25 = 250 cents, or $2.50. The same ten hooks at 4 seconds, which is the shortest length on Seedance and Kling 3, cost $5.00 on Wan 3.0 and 10 x 232 = $23.20 on Seedance 2.5. A 3-second Omni clip at 720p is 38 cents, the same as Wan at 3 seconds.
So for a pure hook test where two seconds is enough, Wan 3.0 is the cheapest row, and the cost grows linearly if you move to longer cuts.
Sources
Related posts
More in Pricing
- What $20 buys: 1,428 seconds at Vidu's from-rate vs Sume clips
$20 / $0.014 = 1,428 s at Vidu Q4's launch rate. On Sume the same $20 buys 21, 10 or 9 fifteen-second clips at 480p, 720p, 1080p (cheapest model each).
- 20 drafts, 5 finals, 1 hero: Omni 7-second clips for $14.63
A three-tier Omni Flash plan for 7-second clips: 20 drafts at 360p, 5 finals at 1080p and 1 hero at 4K cost $14.63 on Sume. Arithmetic by tier.
- 200 support calls of 6 minutes: 20 hours of transcripts, priced
20 hours of support calls cost $10.80 at MAI-Transcribe-2-Streaming's intro price and $12.00 on Sume STT at $0.01 a minute. Job count and limits included.
- 24 field interviews of 25 minutes: 10 hours transcribed
24 field interviews of 25 minutes are 10 hours: $5.40 on MAI-Transcribe-2-Streaming's intro price and $6.00 on Sume STT in 72 jobs of 10 minutes or less.
Written by Sume