Image-to-video API with sound: five-second cost by Sume row
Omni Flash 1.1 (always on), Wan 3.0 (audio), MiniMax H3 and H3 Max (native stereo) take a first frame and render sound. A 5 s clip is $0.19 to $0.94 on Sume.

Four Sume video rows turn a still into a clip that has sound: Gemini Omni Flash 1.1 (native synced audio, always on), Wan 3.0 (listed with audio), and MiniMax H3 and H3 Max (native stereo, no toggle). A 5-second clip costs between $0.19 and $0.94 on the rows in the table, depending on the row and resolution. Sume's Grok Imagine row is the cheap image-to-video option, but the catalog lists it without audio, so it is not in this table.
Sound-capable image-to-video rows at 5 seconds
Rates are the fal list times 1.25, from the Video Router catalog notes. 5-second totals below are unrounded; each job's final price is ceiled to the cent.
| Row | Resolution | Per second | 5 s |
|---|---|---|---|
gemini-omni-flash-1.1 | 360p | $0.0375 | $0.1875 |
gemini-omni-flash-1.1 | 720p | $0.125 | $0.625 |
gemini-omni-flash-1.1 | 1080p | $0.1875 | $0.9375 |
wan-3.0 | 480p | $0.0625 | $0.3125 |
wan-3.0 | 720p | $0.125 | $0.625 |
minimax-h3 | 768p | $0.075 | $0.375 |
minimax-h3-max | 768p | $0.10 | $0.50 |
One request body
Omni accepts 16:9 and 9:16 only, 3 to 10 seconds. Wan 3.0 takes 2 to 30 seconds. H3 and H3 Max take 5 to 15 seconds and produce native 768p, which is not 720p. The first-frame request body is the same on all of them on /v1/videos:
{
"model": "gemini-omni-flash-1.1",
"prompt": "The kettle starts to boil, steam drifts left, kitchen ambience",
"duration": 5,
"resolution": "720p",
"frame_images": [
{
"type": "image_url",
"image_url": { "url": "https://example.com/kettle.jpg" },
"frame_type": "first_frame"
}
]
}Muting
Do not send generate_audio: false to Omni or H3: those models always render sound, and Omni rejects the field. If you need a silent clip for your own music bed, pick a row whose catalog entry lists the toggle, or trim the audio afterwards.
Sources
Related posts
More in Models
- Is Qwen Image 2.1 Turbo on Sume? No: use these three Qwen rows
Sume's image catalog has no Qwen-Image-2.1 Turbo row. It lists Qwen Image, Qwen Image Max and Qwen Image Layered, and one call shows which.
- Long AI video clips on Sume: which rows pass 15 seconds
Only two Sume rows pass 15 seconds in one job: Seedance 2.5 (4-30 s) and Wan 3.0 (2-30 s). All other rows stop at 15 s or less. Read 2026-10-13.
- MAI-Voice-2.1 emotion control vs Sonic's six emotions on Sume
Microsoft lists granular emotion control for MAI-Voice-2.1. On Sume, Sonic 3 rows take six named emotions in generation_config.emotion; Eleven rows have none.
- Sume video model ids in one table (October 2026)
All Sume video ids with resolution and length limits: Seedance, Kling 3, Wan 3.0, H3, H3 Max, Omni, Grok, plus Google Cloud and Tencent rows. Read 2026-10-13.
Written by Sume