News / Model Watch · Official API Launch
HappyHorse 1.1 Goes Global: Full API on Model Studio, Video Editing, and a Climb Past Seedance
Alibaba Cloud has opened HappyHorse 1.1's complete capability set via API on Model Studio — the 快乐马 line's biggest step since it topped the arena at 1.0: a new video-editing mode, integrated audio at no extra cost, a 40% two-week launch discount — and a No. 2 global video-arena rank, ahead of Seedance and Sora.
HappyHorse 1.1 is officially live via API on Alibaba Cloud Model Studio (VentureBeat; Model Studio docs), with four generation modes — text-to-video, image-to-video, subject/reference-to-video, and the new video editing — plus integrated audio at no additional cost. It costs $0.0988/s at 720p and $0.1278/s at 1080p per the OpenRouter listing (served by Alibaba Cloud International), before the 40% sitewide launch discount running two weeks.
What shipped, per official and platform sources
- Full API on Model Studio, globally. Alibaba Cloud's announcement (carried by its official channels and VentureBeat) puts all 1.1 capabilities behind the standard Model Studio API for enterprise and developer use — the "globally available" moment for the line.
- A fourth mode: video editing. 1.0 shipped T2V / I2V / reference-to-video; 1.1 adds editing, which VentureBeat frames as covering "the full spectrum of commercial video needs from ideation through production to post-production."
- Integrated audio at no extra cost — one pass, one price, native audio-video.
- Platform roll call, day one: OpenRouter (
alibaba/happyhorse-1.1, served by Alibaba Cloud International), Together AI (I2V endpoint), WaveSpeedAI (including a video-extend variant), and getimg.ai (1.1 alongside 1.0, commercial rights included). - The ranking story. Per VentureBeat's arena coverage, HappyHorse rises to No. 2 globally as "OpenAI's Sora and ByteDance's Seedance fall away," with Veo 3.1 as the remaining Western rival — and Alibaba's data suggesting HappyHorse outperforms Veo on user-perceived quality.
Pricing (2026)
| Channel | 720p | 1080p | Note |
|---|---|---|---|
| OpenRouter (Alibaba Cloud Intl) | $0.0988/s | $0.1278/s | 3–15s clips; text+images in, video out |
| Model Studio (launch promo) | 40% sitewide discount, first two weeks | official channel; verify live page for promo terms | |
| HappyHorse 1.0 (reference) | ~$0.14/s | ~$0.28/s | third-party list prices at launch (fal); 1.1 undercuts both tiers |
A 10-second 720p clip on 1.1 ≈ $0.99 list, ≈ $0.59 during the launch discount. Verify the Model Studio pricing page before production budgeting.
Get HappyHorse through one OpenAI-compatible endpoint
ChinaModelAPI is an independent relay bringing Chinese frontier models — including the HappyHorse video line — to builders worldwide behind a single OpenAI-compatible API with USDT/USD1 payments. HappyHorse 1.0 is the line's documented baseline today; 1.1 relay routing follows the upstream rollout.
ChinaModelAPI is an independent relay with no official relationship to Alibaba.
Primary sources
- VentureBeat — Alibaba's AI video model rises to No. 2 in global rankings（Model Studio 全量 API + 40% 折扣 + 四模式）
- OpenRouter — alibaba/happyhorse-1.1 定价与规格（$0.0988–$0.1278/s，Alibaba Cloud International 供数）
- Alibaba Cloud 官方 — HappyHorse 1.1 now globally available on Alibaba Cloud（官方公告视频）
- Together AI — HappyHorse 1.1 I2V endpoint
- WaveSpeedAI — HappyHorse 1.1 collection（含 video-extend）
- getimg.ai — HappyHorse 1.1（商业授权 + API）
FAQ (2026)
Officially on the API?
Yes — full capabilities live on Alibaba Cloud Model Studio, plus OpenRouter / Together / WaveSpeed / getimg from day one.
What does it cost?
$0.0988/s (720p) and $0.1278/s (1080p) per OpenRouter, before Model Studio's two-week 40% launch discount. 10s @720p ≈ $0.99 list.
New vs 1.0?
Video editing joins T2V / I2V / reference-to-video; integrated audio at no extra cost; better motion and subject consistency; video-extend on WaveSpeed.
Against Seedance / Kling / Sora?
Per VentureBeat's arena coverage: No. 2 globally — ahead of Seedance and Sora, with Veo 3.1 the remaining Western rival.
Inputs and outputs?
Text + images in (first frame or references); 3–15 second video out, up to 1080p, native audio.
Via ChinaModelAPI?
1.0 is the documented line today; 1.1 relay routing follows the upstream rollout — waitlist gets first notice.