30-second explainer on four video models: cost and how many fit $20
One project, six 5-second shots, three stills, 450 characters of voice, one music bed, one render: total per video model and how many fit a $20 wallet.

A 30-second explainer that mixes images, video, voice and music costs $4.45 on wan-3.0 at 720p and $18.03 on seedance-2.5 at 720p, on Sume list prices. The video model decides almost all of it: the non-video lines add up to $0.6964 in every case. A $20 wallet covers four wan-3.0 versions, five minimax-h3-max versions, and one version on either Seedance model.
The project here is six 5-second shots, three stills for title cards, a 450-character voice-over, one music bed, and one one-minute timeline render to join them. The prices are the Sume rates in the public catalog (read 2026-10-08).
The lines that do not change
Everything except the video shots is the same on all four rows. Three nano-banana-2.1 images at 2K are 3 x $0.15 = $0.4500. The voice-over is 450 characters at $0.0475 per 1,000, so 0.45 x $0.0475 = $0.021375. One Music 1.0 generation is a flat $0.125. The timeline render is $0.10 per output minute, rounded up, so a 30-second result is one billed minute: $0.10.
Fixed lines: $0.4500 + $0.021375 + $0.1250 + $0.1000 = $0.696375.
Per video model
Video is billed per output second at the model's resolution row. Thirty seconds of footage is 6 shots of 5 seconds. Every one of these models accepts a 5-second clip (wan-3.0 takes 2 to 30 s, minimax-h3-max 5 to 15 s, seedance-2 4 to 15 s, seedance-2.5 4 to 30 s).
| Video model | Video line | Project total | Whole projects in $20 | Left over |
|---|---|---|---|---|
| wan-3.0 (720p) | 30 s x $0.1250 = $3.7500 | $4.4464 | 4 | $2.21 |
| minimax-h3-max (768p) | 30 s x $0.1000 = $3.0000 | $3.6964 | 5 | $1.52 |
| seedance-2 (720p) | 30 s x $0.3780 = $11.3400 | $12.0364 | 1 | $7.96 |
| seedance-2.5 (720p) | 30 s x $0.5778 = $17.3340 | $18.0304 | 1 | $1.97 |
What to do with the leftover
The leftover column is where the wallet decision sits. On seedance-2 one project leaves $7.96, which is not enough for a second project but would pay for 4 extra 5-second shots at $1.8900 each. On wan-3.0 the four projects leave $2.21. Decide the model first, then size the wallet in whole projects.
The totals above are list-price estimates for default options. A different image size, a longer voice-over or a second music bed moves only its own line, so keep the line items in your own sheet and change one number at a time.
Reserve before capture
Each paid step reserves its estimate at submit and captures it on success, so the balance you can spend is lower while jobs are running. For the wan-3.0 version, six shots reserve 6 x $0.625 = $3.75 up front if you submit them together, before any of them finishes. A wallet that holds exactly one project total can therefore reject the render, which is the last submit, if the shots are still reserved when you send it.
The practical rule is to keep a margin of one extra shot above the project total, and to submit the render only after the shots complete, because the render needs their finished files as inputs anyway.
Sources
Related posts
More in Pricing
- 30-second video ad on Sume: one clip or two joined, priced
One 30 s Seedance 2.5 clip is $17.33 at 720p; two joined 15 s Seedance 2 Fast clips are $9.17 with the $0.10 Timeline join. Wan 3.0 30 s is $3.75.
- 30 voiceovers of 45 seconds a month: 80 to 86 cents on Sume TTS
A 45-second voiceover is about 562 to 600 characters, or $0.027 to $0.029 on Sume TTS. Thirty a month is 80 to 86 cents. Arithmetic and the rounding caveat.
- 300 UGC videos of 45 seconds: the transcript bill
300 creator videos of 45 seconds are 3.75 hours: $2.03 at MAI-Transcribe-2-Streaming's intro price, $2.25 on Sume STT. Detach and word-timing notes.
- 31-second driver is over Kling Motion Control's 30 s limit
Kling 3.0 Motion Control accepts 1 to 30 seconds at $0.1575 per second. Trim a 31 second driver to 30 and the job costs $4.725.
Written by Sume