HappyHorse 1.1 is a video generation model from Alibaba. It generates short videos from a text prompt, a single starting image, or a set of reference images, with output up to 1080p and durations of 3 to 15 seconds. It is suited for creative content, social media clips, and image-driven animation, and improves on the prior version with stronger prompt adherence, smoother motion, and more consistent characters across frames.
HappyHorse 1.1 is a video generation model from Alibaba. It generates short videos from a text prompt, a single starting image, or a set of reference images, with output up to 1080p and durations of 3 to 15 seconds. It is suited for creative content, social media clips, and image-driven animation, and improves on the prior version with stronger prompt adherence, smoother motion, and more consistent characters across frames.
HappyHorse 1.1 is a video generation model from Alibaba. It generates short videos from a text prompt, a single starting image, or a set of reference images, with output up to 1080p and durations of 3 to 15 seconds. It is suited for creative content, social media clips, and image-driven animation, and improves on the prior version with stronger prompt adherence, smoother motion, and more consistent characters across frames.
HappyHorse 1.0 is a video generation model from Alibaba. It generates short videos from a text prompt, a single starting image, or a set of reference images, with output up to 1080p and durations of 3 to 15 seconds. It is suited for creative content, social media clips, and image-driven animation across a range of aspect ratios.
HappyHorse 1.0 is a video generation model from Alibaba. It generates short videos from a text prompt, a single starting image, or a set of reference images, with output up to 1080p and durations of 3 to 15 seconds. It is suited for creative content, social media clips, and image-driven animation across a range of aspect ratios.
HappyHorse 1.0 is a video generation model from Alibaba. It generates short videos from a text prompt, a single starting image, or a set of reference images, with output up to 1080p and durations of 3 to 15 seconds. It is suited for creative content, social media clips, and image-driven animation across a range of aspect ratios.
Wan 2.7 is a video generation model from Alibaba. It supports text-to-video, image-to-video with first and last frame control, and reference-to-video, where multiple reference images guide the style and content of the generated scene.
Wan 2.7 is a video generation model from Alibaba. It supports text-to-video, image-to-video with first and last frame control, and reference-to-video, where multiple reference images guide the style and content of the generated scene.
Wan 2.7 is a video generation model from Alibaba. It supports text-to-video, image-to-video with first and last frame control, and reference-to-video, where multiple reference images guide the style and content of the generated scene.
Wan2.6-i2v represents a major leap in temporal consistency and narrative depth for image-to-video generation. Building on the Wan2.5 architecture, the 2.6 version introduces an intelligent multi-shot storyboard engine that can transition a single input image into a 15-second cinematic sequence containing multiple cuts. It features a novel "Video-As-Prompt" mechanism and a dedicated MoE (Mixture-of-Experts) backbone that preserves character identity and background stability even during complex camera pans or scene changes. Most notably, it generates studio-grade, synchronized audio (dialogue, foley, and score) natively alongside the visual frames, achieving a level of production readiness previously unseen in open-source models.
Wan2.6-i2v represents a major leap in temporal consistency and narrative depth for image-to-video generation. Building on the Wan2.5 architecture, the 2.6 version introduces an intelligent multi-shot storyboard engine that can transition a single input image into a 15-second cinematic sequence containing multiple cuts. It features a novel "Video-As-Prompt" mechanism and a dedicated MoE (Mixture-of-Experts) backbone that preserves character identity and background stability even during complex camera pans or scene changes. Most notably, it generates studio-grade, synchronized audio (dialogue, foley, and score) natively alongside the visual frames, achieving a level of production readiness previously unseen in open-source models.
Wan2.6-r2v is a revolutionary Reference-to-Video foundation model from the Tongyi Wanxiang series. It allows users to upload "Reference Videos" (containing specific characters, voices, or objects) and use text prompts to generate entirely new cinematic scenes that preserve the original subject's appearance and voice with high fidelity. As China's first multimodal reference model, it supports native audio-visual synchronization, enabling consistent character dialogue and sound effects across multiple shots. It is specifically optimized for professional short-drama production, allowing creators to "star" in AI-generated videos while maintaining a consistent visual identity and auditory persona.
Wan2.6-T2V is the latest flagship text-to-video model from Alibaba's Tongyi Wanx series, optimized for professional cinematic production. Its most significant breakthrough over previous versions is the support for 15-second high-definition video generation and intelligent multi-shot storytelling. The model features enhanced spatio-temporal consistency, ensuring that characters, costumes, and environments remain stable across complex camera movements and scene transitions. Additionally, it natively integrates Audio-Visual Synchronization, automatically generating realistic sound effects (SFX) or syncing provided audio directly with the visual flow.
Wan2.5-i2v-preview is a foundational multimodal video generation model from Alibaba's Tongyi Wanxiang series. It serves as a research preview that pioneered native audio-visual synchronization in the Wan series. The model can transform a single static image into a high-definition video (up to 1080p) while simultaneously generating matching sound effects, ambient noise, or speech. It excels in prompt adherence and temporal consistency, making it a powerful tool for cinematic storytelling and digital human animation (lip-sync support).