Alibaba AI video model climbs to No. 2 as Sora collapses and Seedance stalls

Alibaba Cloud unveiled HappyHorse 1.1 on Sunday, a major upgrade to its Alibaba AI video model that promises production‑grade video synthesis for enterprises. The model is now live on Alibaba Cloud Model Studio with full API access and a ↑ 40% launch discount for the first two weeks.

Alibaba AI video model ranks second globally

Independent benchmark Arena.ai places the model at No. 2 across text‑to‑video, image‑to‑video and hybrid leaderboards, edging out Google’s Veo‑3.1 by 69 points in text‑to‑video and xAI’s Grok‑Imagine‑Video by 23 points in image‑to‑video. In an Elo‑style system, such gaps reflect consistent human preference, not statistical noise.

The 1.1 upgrade adds multi‑image reference (R2V), allowing consistent character identity across frames—a long‑standing hurdle for AI video. Motion modeling is tightened, facial artifacts like “oiliness” and “over‑sharpening” are removed, and lip‑sync drift is eliminated, delivering “zero‑drift” audio‑visual alignment.

“We are delivering a unified 15‑billion‑parameter transformer that handles text, image, video and audio in a single pass,” said Dr. Feifei Li, Alibaba Cloud CTO, in a statement to Reuters.

While OpenAI’s Sora was discontinued after spiralling costs of ↓ $1 million per day and dwindling users, and ByteDance’s Seedance 2.0 hit a legal wall over Bloomberg reports of copyright claims, Alibaba faces fewer rivals. Google’s Veo remains the primary Western contender, but Alibaba’s pricing—roughly $1.82 per 10‑second 720p clip at 1.0—could drop further with the promotion, opening the market to mid‑size agencies.

The model rides on Alibaba’s $52.7 billion cloud buildout, with new data centers in France and Japan reducing latency and meeting European data‑sovereignty rules that emerged after the pandemic era. Yet the Pentagon’s recent listing of Alibaba as a Chinese military‑linked firm adds a compliance layer for U.S.‑exposed customers.

Enterprises should watch adoption metrics, third‑party platform updates and any shifts in regional certification as the AI video market consolidates around a single, well‑funded contender.