Wan 2.7 Uncensored

• 🆕 Text-to-video model from Alibaba's Wan 2.7 generation
• 🏢 Built by Alibaba, served through its Model Studio API
• 👁️ Photorealistic output with strong cross-frame subject coherence
• 🔧 Wan family spans text-, image-, and reference-to-video
• 🔒 Uncensored build relaxing standard Wan content moderation
• 🌐 Released June 2026, newest in this Wan family line

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 2.7 Uncensored is a text-to-video model in Alibaba's Wan family, offered here as an uncensored variant of the underlying Wan 2.7 generation. The standard line is documented by Alibaba Cloud's Model Studio, which exposes Wan endpoints for both text-to-video and image-to-video generation through its DashScope SDK. Within the catalog, Wan is positioned for photorealistic generation with strong subject coherence across frames.

This text-to-video build sits alongside the companion Wan 2.7 Uncensored image-to-video model, which conditions on an input frame rather than a prompt, and the base Wan 2.7 text-to-video endpoint from the same generation. It follows the earlier Wan 2.6 release in the family's lineage, making 2.7 the newer of the two text-to-video generations represented here.

The uncensored designation distinguishes this build from the standard Wan endpoints, which apply input and output content moderation as part of the managed API flow. Beyond that relaxation, it shares the broader generation's video-generation focus.

For practical use, confirm specific resolution limits, clip-length ceilings, licensing, and commercial terms against Alibaba's official Wan and Model Studio documentation before production deployment, since these vary by endpoint and region.