Cinematic 1080p Video
HappyHorse 1.1 generates full 1080p video with rich detail and stable motion — a major leap for an open-source model.
AI video model
HappyHorse 1.1 — The #1 Open-Source AI Video Model, Live on Viberot
HappyHorse 1.1 is a unified Transformer that produces cinematic 1080p video from a text prompt, a single image, or up to nine reference images. It topped the Artificial Analysis AI Video Arena, surpassing Seedance 2.0 — and you can generate with it on Viberot right now.
Made with HappyHorse 1.12 examples
01 — How it works
How to Use HappyHorse 1.1 on Viberot
From prompt to finished video
01
Select HappyHorse 1.1
Pick HappyHorse 1.1 from the Viberot video model selector. No setup, no GPU, no install.
02
Write Your Prompt
Describe the scene, subject, camera movement and mood. HappyHorse 1.1 accepts prompts up to roughly 5,000 characters.
03
Choose Your Input Mode
Go text-to-video, upload one image to animate it, or supply up to nine reference images to lock characters and style. Then pick 720p or 1080p and a duration from 3 to 15 seconds.
04
Generate & Download
Hit Generate and Viberot streams the finished video straight to your library, watermark-free on paid plans.
HappyHorse 1.1 runs on the same workflow as every other model on Viberot — select the model, write a prompt, pick a duration, and generate.
02 — Features
What Makes HappyHorse 1.1 Special
HappyHorse 1.1 replaces multi-stream complexity with a single self-attention Transformer. The result is stable motion, strong prompt adherence, and three distinct ways to drive a shot — text, a single image, or a set of reference images.
Up to 9 Reference Images
Reference-to-video accepts up to nine images, so you can lock characters, props and style across a shot instead of describing them in words.
3 to 15 Seconds
Pick any duration on a continuous slider rather than a handful of fixed presets, and pay only for the seconds you generate.
#1 on AI Video Arena
HappyHorse topped the Artificial Analysis Video Arena in real user blind tests, surpassing Seedance 2.0.
03 — Built for creators
HappyHorse 1.1 on Viberot — Open-Source Quality, Studio Workflow
HappyHorse 1.1 is open-source, but running it well requires serious GPUs. Viberot hosts it for you with zero setup and the same credit balance you already use for Veo, Seedance and Kling.
- No GPUs, no installs — use HappyHorse 1.1 straight from the browser
- Unified credits across every AI video model on Viberot
- Commercial-use license preserved for paid-plan output
- Per-second pricing — a 4-second clip costs less than a 15-second one
04 — Why Viberot
HappyHorse 1.1 vs Other AI Video Models
HappyHorse is the first open-source model to rank #1 on the public AI Video Arena. Here is how it stacks up against the current Viberot lineup.
#1 on AI Video Arena
Blind-test Elo score above Seedance 2.0 and Veo 3.1 Fast at the time of this page.
Fully Open Source
Base model, distilled model, super-resolution module and inference code are released with commercial rights.
Nine Reference Images
More reference slots than most models here — useful for keeping a character consistent across a series of shots.
Continuous Duration
Any length from 3 to 15 seconds, billed per second, instead of fixed 5s / 10s tiers.
Single-Stream Transformer
A self-attention-only architecture — simpler, faster to iterate on, easier to fine-tune.
Commercial Use
Output on paid Viberot plans is watermark-free and safe for commercial projects.
06 — FAQ
HappyHorse 1.1 FAQ
Common questions
What is HappyHorse 1.1?
HappyHorse 1.1 is an open-source unified Transformer that generates 1080p video from text prompts, a single image, or up to nine reference images. It topped the Artificial Analysis AI Video Arena.
Does HappyHorse 1.1 generate audio?
No. HappyHorse 1.1 produces video only — there is no audio track and no lip sync. If you need synchronized native audio, use Seedance 1.5 Pro, Seedance 2 Mini or Kling 3.0 on Viberot.
How much does it cost?
HappyHorse 1.1 is billed per second of video: 56 credits per second at 720p and 73 credits per second at 1080p. A 5-second 720p clip costs 280 credits.
What resolutions and durations are supported?
720p and 1080p, with any duration from 3 to 15 seconds selected on a continuous slider.
Is image-to-video supported?
Yes. Upload a single image to animate it, or supply up to nine reference images to guide characters and style. Note that in single-image mode the output follows your image's aspect ratio, so the aspect ratio selector is hidden.
Which languages can I prompt in?
HappyHorse 1.1 has strong multilingual prompt understanding. Prompts are capped at roughly 5,000 non-Chinese characters or 2,500 Chinese characters.
When should I use something else?
Pick Seedance 2 Mini or Kling 3.0 when you need native audio, Grok Imagine 1.5 when you want the cheapest possible draft, and Veo 3.1 Quality when you need maximum fidelity.
More like this
Try these next.
Generate with HappyHorse 1.1 on Viberot
HappyHorse 1.1 is the #1 open-source AI video model, and it is live on Viberot today. Create a free account and generate your first 1080p clip in under a minute.