Cheapest 10-second image-to-video clip: price by model on Sume
A 10 second image-to-video clip ranges from $0.125 on Grok to $1.40 on Kling 3. Eight Sume catalog rows at list x 1.25, with the limits that decide the pick.

The short answer
The cheapest 10-second image-to-video clip on Sume is Grok Imagine Video 1.5 at $0.125 (10 seconds at $0.0125). The next rows are Gemini Omni Flash 1.1 at 360p ($0.375) and MiniMax H3, MiniMax H3 Max or Wan 3.0 at 480p ($0.625 each). Kling 3 with no audio is $1.40. Every price below is provider list times 1.25.
Ten seconds, eight rows
The table sorts the same 10-second duration across the catalog rows that accept a first-frame image. Only the rows with a documented per-second rate are listed.
| Model | Resolution | Per second | 10 seconds |
|---|---|---|---|
| grok-imagine-video-1.5 | 480p or 720p | $0.0125 | $0.125 |
| gemini-omni-flash-1.1 | 360p | $0.0375 | $0.375 |
| minimax-h3 | 480p | $0.0625 | $0.625 |
| minimax-h3-max | 480p | $0.0625 | $0.625 |
| wan-3.0 | 480p | $0.0625 | $0.625 |
| minimax-h3 | 768p | $0.075 | $0.75 |
| gemini-omni-flash-1.1 | 720p | $0.125 | $1.25 |
| kling-3 | 720p or 1080p, silent | $0.14 | $1.40 |
Which of these accept 10 seconds
Price is not the only filter at 10 seconds. Omni Flash 1.1 accepts 3 to 10 seconds, so 10 is its ceiling. Grok takes 4 to 15 seconds, Wan 3.0 takes 2 to 30 seconds, and MiniMax H3 takes 5 to 15. All four fit a 10-second request. A 10-second clip also needs at least 10 seconds of story: if the picture only supports a 3-second move, the cheap model will pad it with drift.
What the cheap rows give up
Each row has a trade. Grok is silent and has no last frame. Omni always generates native audio, so its price already includes sound. H3 always generates audio too. Wan 3.0 is the long-clip row. Kling 3 is $0.21 a second with audio, so a 10-second clip is $2.10, and it is the only one in this table with both a first and a last frame at 720p and 1080p for the same price.
The same request for every row
Pick the model with model, give the picture as a first_frame, and set duration to 10. The call below is a Grok request; swap the model string to compare.
curl -X POST https://api.sume.com/v1/videos \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: tenseconds-001" \
-d '{
"model": "grok-imagine-video-1.5",
"prompt": "Slow push-in, steam rising from the cup",
"duration": 10,
"resolution": "720p",
"frame_images": [{
"type": "image_url",
"image_url": {"url": "https://example.com/cup.png"},
"frame_type": "first_frame"
}]
}'Check the number before a batch
Sume rounds the billed amount for a job, so check the cost on a 5-second test before a batch. The submit call returns 202 and a polling_url; the final price is whatever the completed job reports. For the full per-model parameter list, read the Videos API docs.
Sources
Related posts
More in Pricing
- Cheapest 4K image on Sume: GPT Image 2.5 $0.014 vs Nano Banana
GPT Image 2.5 renders 3840x2160 for $0.014 at low and $0.0325 at medium on Sume; Nano Banana 2 4K is $0.20 and Pro is $0.375. What the 4K price rows include.
- Cheapest draft image on Sume: GPT Image 2.5 low at 0.74 cents
The cheapest image you can draft on Sume is GPT Image 2.5 at low quality, $0.0074, about a third of Grok Imagine and a tenth of Nano Banana 2 at 0.5K.
- Cheapest image edit models on Sume: price per billed image
Six Sume image ids that take references cost under $0.06 billed per image. A ranked table from the catalog list prices, with caveats on n and quality.
- Cheapest Seedance 2.5 test clip: 4 s at 480p is $1.08
The shortest Seedance 2.5 job on Sume is 4 s. At 480p that is $1.08; 720p is $2.31; 1080p is $5.69. Wan 3.0 starts at 2 s and 13 cents. Test routine inside.
Written by Sume