Alibaba released the Happy Horse video generator — it took first place in global benchmarks

Alibaba quietly launched the Happy Horse video generator — and it immediately took first place in global benchmarks, surpassing Sora, Runway Gen-4, and Kling. For Western AI companies, this is an unwelcome surprise: first place in one of the most technically complex segments is now held by China.

Stealth Market Entry

Alibaba made no grand announcements: Happy Horse appeared without a press conference or marketing campaign. The model launched "under stealth" — quietly, almost unnoticed. Nevertheless, within a few days it spread through developer communities and made headlines in tech publications when results from the first independent tests proved convincing.

This approach — minimal noise at release, followed by organic momentum — has become the hallmark of Chinese AI labs. DeepSeek R1 arrived in January 2025 without press releases, Qwen launched quietly, Wan Video emerged almost unnoticed — but each time the community itself drove the wave. Happy Horse follows the same logic.

The name, apparently, references 快马 (kuài mǎ) — "fast horse" in Chinese. A symbolic name for a model that claims to outpace all competitors.

What the Benchmarks Show

According to Bloomberg Tech, Happy Horse ranked first on key public benchmarks for video generation. The evaluation was based on several parameters:

• Physical plausibility of object and character movement
• Accuracy in following text instructions (text-to-video alignment)
• Consistency of characters over several seconds
• Visual detail and image clarity
• Generation speed at comparable quality

The model surpassed Sora from OpenAI, Runway Gen-4, and Kling from Kuaishou — the three strongest players who previously dominated the top positions in global rankings. This is the first case where a Chinese AI model holds absolute first place specifically in video generation — unlike text breakthroughs (DeepSeek) and multimodality (Qwen), where leadership has shifted more frequently. Independent testers from the AI community noted that Happy Horse is particularly strong in dynamic scenes: water movement, fabric in wind, facial expressions — traditionally difficult tasks for video models.

Why Video Became the Main Bet

Video generation is the new main frontier in the AI company race. If 2024 saw the main battle in text models and image generators, by 2026 all major labs have switched to video: the computational barrier is higher, monetization potential is greater, competitive advantage is harder to copy. Alibaba is no newcomer to this race. The company actively develops the Qwen series, demonstrated strong results in multimodal tasks, and its cloud division Alibaba Cloud competes with Azure and GCP in the enterprise AI infrastructure segment. Happy Horse is a logical continuation of this strategy, aimed at capturing the market for tools used by content creators, studios, and advertising agencies.

"Alibaba has given

China the crown in video generation," — Bloomberg Tech, April 2026.

For Western players, this means intensifying competition. OpenAI and Runway are actively building infrastructure and investing in video generation — but both now find themselves behind first place.

What This Means

First place on benchmarks is a signal for the entire market: the AI race in video has ceased to be exclusively a Western story. For business, this means the emergence of another strong — and potentially cheaper — tool in the stack for advertising, educational content, and video production automation. Alibaba traditionally releases models at competitive prices, which could shift pricing pressure in the AI Video API segment.