We put Seedance 2.5, MiniMax H3, Grok Imagine Video 1.5 and HappyHorse 1.1 against each other across three image to video scenes, all run on EmpirioLabs. Every model animated the exact same starting image, one generation each, so any difference you see is the model, not a different starting picture.
Watch the four way comparison
Specs at a glance
| Seedance 2.5 | MiniMax H3 | Grok Imagine Video 1.5 | HappyHorse 1.1 | |
|---|---|---|---|---|
| Maker | ByteDance | MiniMax | xAI | Alibaba Cloud |
| Inputs | Text, images, video and audio | Text, images, video and audio | Text and images | Text and images |
| Highest resolution | 1080p | 2K | 1080p | 1080p |
| Clip length | 4 to 30 seconds | 4 to 15 seconds | 1 to 15 seconds | 3 to 15 seconds |
| Vertical 9:16 output | Yes | Yes | Yes | Yes |
| Price per second as tested | $0.462 at 720p | $0.18 at 768p | $0.168 at 720p | $0.14 at 720p |
Some models add a small fee per input image; each model page lists its full pricing.
How we ran it
For each scene we first generated one starting image with Seedream 5.0 Pro at 1584 by 2816, a vertical 9:16 frame. Then every model animated that same image, one generation each, 5 seconds per clip at 720p (MiniMax H3 at 768p, its lowest setting), in 9:16, with otherwise default settings. The three scenes were a red fox diving headfirst into deep snow, a black panther leaping between rainforest branches, and a submarine bursting up through Arctic sea ice. Each model got the same motion prompt per scene, asking for a single continuous shot with no cuts.
What to look for
Because every model starts from the same frame, this comparison is about motion. Watch how the fox dives and the snow bursts around it, how the panther's body and tail move through the leap, and how the ice cracks and slides off the rising hull. Look at how closely each clip follows the motion prompt and how stable the subject stays from frame to frame. We are not declaring a winner. Run the clip, see which look fits the kind of video you make, and pick from there.
Generate any of these on EmpirioLabs
All four models serve the same image to video endpoint: send a starting image and a motion prompt, then swap the model id to compare.
curl https://api.empiriolabs.ai/v1/videos/generations \
-H "Authorization: Bearer $EMPIRIOLABS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "seedance-2-5",
"image": "https://example.com/your-vertical-frame.png",
"prompt": "The fox dives headfirst into the deep snow, single continuous shot, no cuts.",
"resolution": "720p",
"duration": 5
}'Change model to minimax-h3, grok-imagine-video-1-5 or happyhorse-1-1 and send the same image and prompt again. For MiniMax H3, set resolution to 768p or 2k. You can also try them in the playground.
Frequently asked questions
Which video models were tested?
Seedance 2.5 from ByteDance, MiniMax H3 from MiniMax, Grok Imagine Video 1.5 from xAI and HappyHorse 1.1 from Alibaba Cloud. All four run on EmpirioLabs.
Did every model animate the same image?
Yes. We generated one starting frame per scene with Seedream 5.0 Pro, then gave that same frame to all four models, so the comparison isolates the motion from the image.
What settings were used?
5 seconds per clip at 720p (MiniMax H3 at 768p), in 9:16, otherwise default settings, one generation each, no edits or retries. Each model got the same motion prompt per scene.
Can I generate vertical video?
Yes. All four models produce 9:16 video. Set aspect_ratio to 9:16 on MiniMax H3, Grok Imagine Video 1.5 and HappyHorse 1.1. With a starting image, Seedance 2.5 takes adaptive (its default) and follows the image's shape.
How do I pay for it?
Pay as you go, or choose a plan with weekly allowances. Each model page lists its per-second price by resolution.
