What Is HappyHorse 1.1? Alibaba's Video Model Explained

HappyHorse 1.1 is an AI video model from Alibaba, released on June 22, 2026 as the successor to HappyHorse 1.0. On VerseIn, it turns a text prompt, an image or a set of reference images into a silent clip of 3 to 15 seconds at 720p or 1080p. Specs and claims in this explainer were collected from Alibaba Cloud's documentation and VerseIn's HappyHorse 1.1 page on October 22, 2026. We have not generated clips with the model for this article.

Key Takeaways

• Alibaba calls HappyHorse 1.1 a "comprehensive upgrade over version 1.0" and lists changes to motion, character consistency, instruction following, visual texture and audio.
• On VerseIn, the model offers text to video, image to video and reference to video, with clips of 3 to 15 seconds, 720p or 1080p, nine aspect ratios and prompts of up to 2,500 characters.
• VerseIn lists HappyHorse 1.1 output as video only. Its FAQ says you can add voice-over or music afterwards on VerseIn.
• Alibaba's own guidance points to HappyHorse for clips of 15 seconds or shorter and to Wan 3.0 for clips of up to 30 seconds with multimodal references.

What is HappyHorse 1.1?

HappyHorse is Alibaba's family of video generation models. The first version, HappyHorse 1.0, launched in April 2026, according to Alibaba Cloud's list of newly released models. Version 1.1 followed in June.

In its release announcement, Alibaba Cloud says the model targets short drama, e-commerce ads, brand marketing and game CG. The HappyHorse campaign page frames it around cinematic creative work for advertising and online stores.

Developers reach the model through Alibaba Cloud Model Studio. Alibaba's video generation docs give one model ID per mode:

• happyhorse-1.1-t2v: text to video
• happyhorse-1.1-i2v: image to video from a first frame
• happyhorse-1.1-r2v: reference to video

Video editing in that API still runs on a 1.0 model.

VerseIn's model page describes HappyHorse 1.1 as a model that "improves on HappyHorse-1.0 with better prompt understanding." Alibaba's announcement lists a related area, instruction following.

What changed from HappyHorse 1.0, according to Alibaba

Alibaba names five areas of change in the 1.1 announcement:

• Motion: how people and objects move through the scene
• Character consistency: keeping the same person or character recognizable from frame to frame
• Instruction following: how closely the clip tracks what the prompt asks for
• Visual texture: surface detail in the picture
• Audio: sound generated with the video in Alibaba's API

A separate Alibaba Cloud blog post adds one detail for reference to video: the update improves how the model combines several reference images in one generation.

This post does not cover benchmark figures, so treat these points as Alibaba's description of the update. Running your own prompt is a direct way to judge whether they matter for your project. The audio change applies to Alibaba's API. On VerseIn, the model produces video only.

HappyHorse 1.1 specs on VerseIn

The table lists the values shown on VerseIn's model page. Alibaba's API documents a few extra options, such as 480p output and audio, that the VerseIn page does not list.

Aspect ratios and prompt length

The nine ratios include 4:5 and 5:4, which VerseIn's Wan 3.0 and PixVerse pages do not list. A 4:5 frame suits portrait feed posts. The 2,500-character prompt has room for subject, setting, camera movement and mood.

Here is an example input for a 4:5 product clip, written to show that structure: "A ceramic coffee mug on a wooden counter in soft morning light, steam rising from the cup, slow push-in camera, warm and calm mood."

How references work

Alibaba's reference-to-video API docs say the mode accepts 1 to 9 reference images. The prompt cites them as [Image 1], [Image 2] and so on.

Image to video works differently. Alibaba's image-to-video docs say that mode takes exactly one image, used as the first frame.

VerseIn Studio currently accepts up to 9 reference images, matching Alibaba's documented API maximum.

Why is HappyHorse 1.1 silent on VerseIn?

VerseIn's model page lists the output as "Video only," and its FAQ says the model generates silent video. The same FAQ adds: "You can add voice-over or music afterwards on VerseIn."

VerseIn has a separate tool for each kind of sound:

• AI Voice Over turns a script into narration for a product demo or explainer.
• AI Music Generator creates a background track in the mood you describe.
• AI Sound Effects makes sounds such as footsteps, doors or ambient noise.

Because the audio is generated separately, you can make several voice-over or music versions without regenerating the video. If you would rather get sound in the same generation, our guide to AI video generators with audio covers models that output both.

HappyHorse 1.1 vs Wan 3.0 on VerseIn

Wan 3.0 is another Alibaba video model on VerseIn, alongside Wan 3.0 Prime. The table compares HappyHorse 1.1 with Wan 3.0 using the values on each model page.

Alibaba's video generation docs split the two by clip length. They say HappyHorse fits clips of 15 seconds or shorter at up to 1080p, while Wan 3.0 fits clips of up to 30 seconds with multimodal references.

Suggestions based on those specs and that guidance:

• Choose HappyHorse 1.1 when you need a 4:5 or 5:4 frame, plan to add your own voice-over and music anyway, or want text to video, a mode its VerseIn page lists and the Wan 3.0 page does not.
• Choose Wan 3.0 when the clip runs past 15 seconds, you want to set both the first and last frame, or you want sound from the same generation.
• Still undecided? Write one brief, make a short clip from it with each model, and keep the one that fits your project.

For more on Wan 3.0, see VerseIn's Wan 3.0 page. VerseIn's AI Video Generator page gathers the rest of its video models in one place.

How to use HappyHorse 1.1 on VerseIn

1. Open the model. Go to the HappyHorse 1.1 page and open the studio from there.
2. Set up the clip. Write your prompt, or add an image if you want to work from a picture. Pick a length, a quality level and an aspect ratio for the clip.
3. Generate. Review the credit total shown before you confirm, then generate. The finished clip is saved to your Media library.
4. Add sound. If the clip needs audio, create narration, music or sound effects with VerseIn's audio tools to go with it.

For other recent models, see our explainers on PixVerse C1 and Seedance 2.5.

FAQ

Does HappyHorse 1.1 generate audio on VerseIn?

No. VerseIn lists it as video only. Alibaba documents audio for its own API, which VerseIn does not offer.

How long can a HappyHorse 1.1 video be?

On VerseIn, you can choose a length from 3 to 15 seconds. For longer clips, Alibaba points to Wan 3.0, which runs up to 30 seconds on VerseIn.

How many reference images can I use?

Up to 9 reference images can be added in VerseIn Studio. Alibaba's API documents the same maximum.

How many credits does HappyHorse 1.1 use?

Credits are used whenever you generate, and the amount depends on the tool, model and settings. You'll always see the total cost before you confirm. If a generation fails because of a processing error, a timeout or a content safety check, the credits are returned automatically. New accounts get free sign-up credits, and the pricing page lists the plans.