GPT-6 Sol & Claude Opus 5.5 are liveGPT-6 Sol at half the price of 5.6 Sol

HappyHorse 1.0 API

AI Video

HappyHorse 1.0 is the first-generation HappyHorse video family on OmniaKey, published as one card with a model id per operation: happyhorse-1.0-t2v for text to video, happyhorse-1.0-i2v for a first frame, and happyhorse-1.0-r2v for reference-driven shots.

Alibaba

HappyHorse 1.0

Mode
Prompt0 / 12000
Duration
Resolution
Aspect ratio

Estimated $0.56$0.112/s × 5s

Your result will appear here once you run the model.

01

Where it fits

HappyHorse 1.0 is the first-generation HappyHorse video family on OmniaKey, published as one card with a model id per operation: happyhorse-1.0-t2v for text to video, happyhorse-1.0-i2v for a first frame, and happyhorse-1.0-r2v for reference-driven shots.

Use HappyHorse 1.0 to reproduce an existing 1.0-based pipeline; HappyHorse 1.1 undercuts it at both 720P and 1080P, and adds a 480P tier.

02

What it does

Text to video

happyhorse-1.0-t2v

Generate a short video clip from a text prompt.

Image to video

happyhorse-1.0-i2v

Animate a still image into a short video clip.

Reference to video

happyhorse-1.0-r2v

Guide a clip with reference media to keep a subject, style or motion.

03

In depth

HappyHorse 1.0 is the first generation of Alibaba's HappyHorse video family in this catalog. It ships as one card with three ids — happyhorse-1.0-t2v, happyhorse-1.0-i2v and happyhorse-1.0-r2v — matching text, first-frame and reference work respectively, and it prices two resolutions, 720P and 1080P.

There is no 480P entry on this generation, and both of its tiers are listed above the equivalent HappyHorse 1.1 tiers. That pairing is what makes 1.0 a compatibility choice rather than a default: it exists for pipelines whose prompts, crops and grade were tuned against the older look.

Our 14 September 2026 smoke run put one job of each kind through it at 720P, and each came back as an MP4 of 5.16 seconds. Those three files, and a portrait-size comparison against Wan 3.0 from October, are recorded below.

Re-running a tuned 1.0 pipeline

If a prompt library, a crop set and a colour grade were built around 1.0 output, moving them to 1.1 changes the result in ways a client will notice. Keep the old id for those reprints and keep the new work on the newer family.

Matching an older campaign

When a seasonal campaign has to sit beside last year's films, the same generator is the cheapest way to keep the look consistent. Render a reference clip first, then compare it with the archived file before committing the batch.

First frames with their own shape

The i2v id follows the source image: our 5-second test returned 1324x696 from the still we sent, not a letterboxed 16:9 frame. Compose the still in the final shape and leave the ratio out of the request.

04

Hands-on notes

  • Our 14 September 2026 smoke run rendered a 5-second 720p HappyHorse 1.0 text-to-video job as a 1280x720 MP4 running 5.16 seconds.

  • The HappyHorse 1.0 image-to-video job took a first frame and returned a 1324x696 MP4 of 5.16 seconds, following the source image's shape.

  • HappyHorse 1.0 reference-to-video returned a 1280x720 MP4 of 5.16 seconds from one reference image.

  • A 9:16 request at 480P on 25 October 2026 came back from HappyHorse 1.0 as 480x848, slightly narrower than the 480x854 that Wan 3.0 returned.

05

Limits to plan around

  • No 480P tier is priced on this generation, so there is no cheap draft stage inside it; testing a prompt here costs the same per second as finishing it.
  • Both priced resolutions sit above the HappyHorse 1.1 figures, which means the older model is also the dearer one at 720P and 1080P.
  • Three ids have to be handled separately: text, first frame and reference each take a different model name, and a mixed batch needs the id swapped per job.
  • The family stops at 1080P, so a 4K master cannot come out of this id at any rate.

06

Compare nearby models

ModelBest forOperationsListed price
HappyHorse 1.0This page
happyhorse-1.0
The older generation, kept for pipelines already tuned to its look.Text to video · Image to video · Reference to video
$0.112–$0.192
per second
HappyHorse 1.1
happyhorse-1.1
Cheaper at both shared resolutions, adds 480P, and won two of three prompts against Kling.Text to video · Image to video · Reference to video
$0.042–$0.108
per second
Wan 3.0
wan3.0-video
Thirty-second takes, three resolutions and a reference operation at a lower rate.Text to video · Image to video · Reference to video
$0.04–$0.16
per second

08

Call it

Media models run asynchronously. Create a task, poll it until it finishes, then download the signed asset URL.

Create, poll, download
# 1. Create the job. 202 Accepted; the body carries the task id,
#    the pre-charged price, and a poll_url. Reuse one Idempotency-Key
#    across retries so a timeout never starts a second paid job.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks \
  -H "Authorization: Bearer $OMNIAKEY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: happyhorse-1.0-001" \
  -d '{
    "model": "happyhorse-1.0-t2v",
    "type": "video",
    "operation": "text_to_video",
    "input": {"prompt": "a paper boat on a quiet lake at sunset"},
    "parameters": {"duration": 6, "resolution": "720p"}
  }'

# 2. The operation picks the model id on this route:
#      happyhorse-1.0-t2v     prompt only
#      happyhorse-1.0-i2v     media[] type "first_frame"
#      happyhorse-1.0-r2v     media[] type "reference_image"

# 3. Poll GET /v1/media/tasks/{id} until status is
#    succeeded | failed | canceled. Progress carries percent and stage.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks/TASK_ID \
  -H "Authorization: Bearer $OMNIAKEY_API_KEY"

# 4. Download assets[0].url before expires_at, or poll again for a
#    fresh signed URL. Failed and canceled jobs are refunded.
curl -sS -o out.mp4 "$(jq -r '.assets[0].url' result.json)"

09

Pricing

Video jobs are billed per second of rendered output, not per token. Failed or canceled jobs are refunded automatically.

Per second of output

$0.112–$0.192per secondSave 20%

720P
$0.112per second
1080P
$0.192per second

More models from Alibaba

Related articles

Frequently asked questions

How do I call HappyHorse 1.0 on OmniaKey?

POST to /v1/media/tasks with the model id, type "video", and operation "text_to_video" — or "image_to_video" with a first frame in input.media[] with type "first_frame". Reuse one Idempotency-Key across retries so a timeout never starts a second paid job.

What does HappyHorse 1.0 cost?

HappyHorse 1.0 bills $0.112–$0.192 per second of rendered video. Creation pre-charges the requested duration up front from your prepaid balance, and failed or canceled jobs are refunded automatically.

Why does the API return a task id instead of the video?

Video renders take from about a minute to several minutes, so OmniaKey runs every video route through one asynchronous lifecycle. Poll GET /v1/media/tasks/{id} until the status is succeeded, then read the signed asset URL.

Should new work start on HappyHorse 1.0?

Not usually. HappyHorse 1.1 is listed below it at both 720P and 1080P and adds a 480P tier, so a fresh project gets more options for less. Reach for 1.0 when an existing prompt set or grade already depends on its output.

What does HappyHorse 1.0 return for a 9:16 request?

A portrait frame at the width that resolution allows, which in our October test was 480x848, a little narrower than the 480x854 that Wan 3.0 produced for the same shape. Check the dimensions after the download rather than assuming a standard size.

Does the image-to-video id letterbox my still?

No. Our 5-second test returned 1324x696 from the still we supplied, so the frame follows the source image rather than being padded to 16:9. Send the still in the shape you intend to ship and leave the ratio parameter out.

Can I mix HappyHorse 1.0 and 1.1 in one batch?

Yes, but expect visible differences between the two generations in motion and colour. Render the same prompt on both first, compare the frames at matching seconds, and only then decide which shots belong to which generation.

How is a HappyHorse job charged?

Per output second at the resolution you choose, taken from balance when the task is created and returned if the task fails. The catalog charges one rate per resolution, and the current figures are in the price table on the model page, so switching between the three operation ids changes the operation rather than the charge.