on april 27, 2026, alibaba’s video-generation model happyhorse 1.0 launched its gray-scale testing phase. professional creators and enterprise customers worldwide can register and use the service on the happyhorse official website (www.happyhorse.cn) and the alibaba cloud bailian platform, while general users can experience it through the qwen app. on the official website, the standard pricing for 720p video generation is rmb 0.9 per second.
happyhorse 1.0 is built on a native multimodal architecture and employs a joint audio–video generation approach, catering to content-production scenarios such as advertising, e-commerce, short dramas, and social-media creativity, and offering integrated creation capabilities that span from intelligent generation to editing. the happyhorse official website is a professional, all-in-one ai video-creation platform; new users receive a free quota upon registration, with standard pricing of rmb 0.9 per second for 720p videos and rmb 1.6 per second for 1080p videos. for professional members, after applying time-limited discounts, the monthly subscription rates drop to rmb 0.44 per second for 720p and rmb 0.78 per second for 1080p.
happyhorse 1.0 comprises two core functions: video generation and video editing. video generation covers mainstream capabilities such as text-to-video, image-to-video, and multi-image–input video generation, while video editing enables flexible secondary creation of videos. the model supports multi-camera storytelling up to 15 seconds, adaptive aspect-ratio rendering, and 1080p super-resolution output.
the model excels in aspects such as visual texture and lighting effects, smooth camera movements and transitions, and lifelike human figures. whether it’s skin texture or smoke, water, and fog, the results achieve cinematic quality. it particularly excels at capturing medium- and close-up shots with large apertures, shallow depth of field, and strong atmospheric mood, supporting various camera techniques such as zooming in, zooming out, and depth-of-field adjustments. notably, the model also features audio–video synchronization, capable of generating everything from tense dialogue exchanges to ambient sounds like raindrops hitting windowpanes. moreover, it effortlessly reproduces a wide range of styles, from ink-wash meticulous brushwork to origami and clay stop-motion animation.
during the gray-scale testing phase, happyhorse 1.0’s model capabilities continue to be iteratively upgraded. agent platforms such as alibaba wukong, mulerun, and jvs claw have also been integrated. the happyhorse official website is currently running the “super creator · the first 100” campaign, warmly inviting aigc creators from home and abroad to join. users can sign up by completing a questionnaire on the official website.
