Happy Horse 1.1 is Alibaba's video model: up to 15-second 1080p clips with sound, from text, an image or up to 9 reference images. We animated five of our own images — watch the real clips below.
Quick answer
Updated
Happy Horse 1.1 (HappyHorse) is Alibaba's AI video model, developed by Alibaba's Token Hub unit and released on 22 June 2026 as an upgrade to HappyHorse 1.0 (April 2026). It makes 3–15-second clips at up to 1080p and 24 fps with native audio — dialogue, ambient sound and effects — from text, a start image or reference images. Alibaba says 1.1 improves motion, character consistency, instruction following and audio-visual sync. On Kenerate, Happy Horse 1.1 runs at 720p or 1080p in five aspect ratios, with up to 9 reference images.
About the model
Facts from the maker's own announcement and repository, linked below.
| Maker | Alibaba Token Hub (ATH) |
|---|---|
| Released | 1.0 beta 28 Apr 2026; 1.1 on 22 Jun 2026 |
| Modes | Text to video, image to video, reference to video |
| Length | 3–15 seconds |
| Resolution | 480p, 720p, 1080p at 24 fps (Alibaba Model Studio) |
| Audio | Native: dialogue, ambience, effects |
| Aspect ratios | 16:9, 9:16, 1:1, 4:3, 3:4 on Kenerate |
| Official site | happyhorse.com; API on Alibaba Cloud Model Studio |
Real clips
Five real image-to-video clips from our own images: 720p, 5 seconds, with the sound the model generated. Press play to hear them.
Talking portrait with a spoken line
Happy Horse 1.1 · image to video · 720p · 5 s with sound
She looks up from the window, smiles and says 'Good morning, everyone!' in a warm voice, soft room tone
Samurai drawing her katana
Happy Horse 1.1 · image to video · 720p · 5 s with sound
Cherry blossoms swirl around her as she slowly draws the katana and looks into the camera, wind and a soft metallic ring
Lisbon tram street
Happy Horse 1.1 · image to video · 720p · 5 s with sound
The yellow tram rolls slowly up the street, people walk past, laundry sways in the breeze, city ambience and a tram bell
Dog catching a frisbee
Happy Horse 1.1 · image to video · 720p · 5 s with sound
The border collie leaps and catches the frisbee, lands and runs toward the camera, happy barking, summer field ambience
Snowy cabin push-in
Happy Horse 1.1 · image to video · 720p · 5 s with sound
Slow push-in toward the glass cabin as snow falls, warm light flickers inside, soft wind and crackling fire sounds
Models
Every version in the generator, and what each is good at.
| Model | Reference images | Max size | Good for | Open |
|---|---|---|---|---|
| Happy Horse 1.1Alibaba | Up to 9 | 720p, 1080p · 3–15 s | Text, image, reference to video; extend | Open |
| Kenerate Video (Happy Horse 1.0)Alibaba | Up to 9 | 720p, 1080p · 3–15 s | Adds video edit | Open |
| Veo 3.1Google | Up to 3 | 720p–4K · 4–8 s | Premium alternative with sound | Open |
From our tests
Alibaba's Happy Horse runs with the provider's own safety checks. For 18+ video as an adult (18+), see the NSFW AI video generator.
How to
Try Happy Horse 1.1 on your own prompt — 25 free credits on sign-up.
In depth
Background, how to prompt it, and when to pick something else.
Happy Horse generates audio with the picture, so describe it: the ambience (city, wind, a fire), effects (a tram bell, a sword ring) and any spoken line in quotes. In our tests the model followed each of these.
Image to video starts from your exact frame — best for animating a finished image. Reference to video takes up to 9 images and keeps characters and products consistent across new scenes.
We only publish real, verified reviews — never invented quotes. Send a line about your result (and whether we may quote you), and it will appear here.
FAQ
6 questions
Alibaba's AI video model, released on 22 June 2026, that makes 3–15-second clips at up to 1080p with native audio from text, an image or reference images.
Alibaba — its Token Hub business unit. The official site is happyhorse.com, with an API on Alibaba Cloud Model Studio.
Yes — Alibaba lists native audio, including dialogue and ambient sound. Every clip we made came with sound.
Text to video, image to video or up to 9 reference images; 3–15 seconds at 720p or 1080p in 16:9, 9:16, 1:1, 4:3 or 3:4.
Alibaba says 1.1 upgrades motion, character consistency, instruction following, visual texture and audio, with better multi-reference consistency.
3 to 15 seconds on Kenerate, and you can extend a clip.
Explore
Generators, design tools, photo edits, models and comparisons — all in one studio, on one credit balance.
Written by the Kenerate AI team
We run these models in the Kenerate generator. Model facts come from the makers' own pages; specs and limits were checked against our code; the examples are our own outputs. Last updated 27 September 2026. Corrections: hello@kenerateai.com.
Ready when you are
25 free credits on sign-up. No watermark.
