Wan 2.7

• 🆕 Alibaba's Wan 2.7 text-to-image model, released April 2026.
• 🔧 Unified image generation and editing via Model Studio API.
• 👁️ Photorealistic outputs with strong detail preservation.
• 💬 Generates images directly from natural-language prompts.
• 🎯 Part of Alibaba's broader Wan generative family.
• 🏢 Built and served via Alibaba Cloud 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 2.7 is the text-to-image entry in Alibaba's Wan generative family, made available in April 2026 through Alibaba Cloud Model Studio. According to the official Model Studio API reference, it generates images from natural-language prompts and also supports instruction-driven image editing through the same endpoint, combining generation and editing in one model interface. Per the catalog description, outputs target photorealistic results with strong detail preservation.

Within Alibaba's lineup, Wan 2.7 succeeds the earlier Wan 2.6 line, whose Model Studio image and video endpoints are documented separately. The official documentation distinguishes the standard model from the higher-tier Wan 2.7 Pro variant, giving users two service levels for static image generation. For verified capabilities and parameters, the Model Studio API reference remains the authoritative source rather than third-party reseller pages.

Within the broader ecosystem, Wan 2.7 sits alongside the video-focused Wan 2.7 models and Alibaba's separate Qwen Image 2 image family, giving users distinct options for static imagery versus motion. Several benchmark-style and pricing claims circulating online come from resellers rather than Alibaba directly, so they are best treated cautiously. The official Model Studio documentation should be consulted for confirmed specifications and supported features.