Wan 2.6

• 🎬 Alibaba's text-to-video model generating photorealistic clips with frame coherence
• 🏢 Built by Alibaba Cloud within the broader Wan family
• 📏 Outputs 720p or 1080p, with 1080p as default
• 🔧 API accepts text prompts plus optional audio input
• 🎯 Focus on subject coherence and motion across frames
• 🆕 Precedes Wan 2.7, the newer text-to-video entry
• 👁️ Companion image-to-video variant animates a first frame
• 📅 Released December 2025 via Alibaba's Model Studio

Alibaba

Alibaba Group is a Chinese multinational technology company founded in 1999 and headquartered in Hangzhou, Zhejiang. Originally built around e-commerce and cloud computing, Alibaba has become one of the most prolific contributors to open-weight AI research,…

Wan 2.6 is a text-to-video model in Alibaba's Wan family, the company's video generation suite developed under Alibaba Cloud's research division. Released in December 2025, it generates clips from text prompts at 720p or 1080p resolution, with 1080p as the API default, and its request schema also accepts an optional audio file to drive generation. It ships alongside a companion Wan 2.6 image-to-video variant; per Alibaba's documentation, the 2.6 image-to-video generation supports first-frame-to-video only.

Within the family, Wan 2.6 follows the earlier Wan 2.5 Preview and Wan 2.2 A14B text-to-video entries from September 2025. Alibaba's documentation positions the line around photorealistic generation with subject coherence across frames.

Wan 2.6 is no longer the newest text-to-video model in the line. Its successor, Wan 2.7, arrived in April 2026, and Alibaba's own documentation recommends Wan 2.7 over 2.6 and earlier for image-to-video work.

On this catalog, Wan 2.6 is the December 2025 text-to-video entry in the Wan family, while Wan 2.7 is the later release in the same line.