One photo, one 30-second AI video: Wan 3.0 costs $1.875 to $7.50
Wan 3.0 runs up to 30 seconds from one first frame on Sume: $1.875 at 480p, $3.75 at 720p, $7.50 at 1080p. When to chain 15 s clips instead.

The short answer
Wan 3.0 can turn one photo into a 30-second clip on Sume, since it accepts 2 to 30 seconds. At 480p that costs $1.875, at 720p $3.75 and at 1080p $7.50 (30 seconds at $0.0625, $0.125 and $0.25). Kling 3 stops at 15 seconds, and Grok Imagine at 15, so for one uncut 30-second shot Wan is the pick.
The two models that reach 30 seconds
Seedance 2.5 also accepts 4 to 30 seconds, so it is the other candidate in the catalog. The Wan prices below are the ones this post quotes; look up Seedance 2.5 on its own rate card before you compare.
Wan 3.0 prices
The cost is linear in seconds, so a shorter test is a fixed fraction of the final.
| Resolution | Per second | 10 s | 30 s |
|---|---|---|---|
| 480p | $0.0625 | $0.625 | $1.875 |
| 720p | $0.125 | $1.25 | $3.75 |
| 1080p | $0.25 | $2.50 | $7.50 |
Request
Send the photo as a first_frame, set duration to 30, and pick a resolution.
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: wan30-001" \
-d '{
"model": "wan-3.0",
"prompt": "Slow dolly through the garden, petals drifting, light shifting",
"duration": 30,
"resolution": "480p",
"frame_images": [{
"type": "image_url",
"image_url": {"url": "https://example.com/garden.png"},
"frame_type": "first_frame"
}]
}'Long shots drift
A single photo gives the model one starting point and 30 seconds to invent. Long shots drift: the subject can change, and the end of the clip may look little like the photo. Test the idea at 5 seconds and 480p, which is $0.3125, before you spend $1.875 or more on the full length.
When to chain instead
If the story is a sequence, two 15-second clips often beat one 30-second clip. Take the last frame of clip one with the video-frames route, use it as the first frame of clip two, and join the results in a Timeline render for $0.10. The price at 480p is the same $1.875 for the two halves, and you get a choice point in the middle.
Sources
Related posts
More in Models
- GPT Image 2.5 can take 2 minutes: submit async, not sync, on Sume
OpenAI says complex GPT Image 2.5 prompts can run up to 2 minutes. Sume's image route blocks only 30 seconds, so send mode async and poll the job.
- GPT Image 2.5 edits through Sume: mask_url and 16 references
Sume's GPT Image 2.5 takes up to 16 references, an optional mask_url and background auto, transparent or opaque. OpenAI: transparent needs png or webp.
- Pocket TTS license: the repo says MIT, not Apache-2.0
Some roundups call Kyutai's Pocket TTS Apache-2.0. Its GitHub page and LICENSE file read MIT-style. How to check a TTS license before you build on it.
- Pocket TTS runs on 2 CPU cores: what ~200 ms first audio means
Kyutai's Pocket TTS lists 100M parameters, 2 CPU cores and ~200 ms to first audio. Whether that matters for a video voiceover, and a hosted TTS job's numbers.
Written by Sume