alibaba/wan-v2-6 models

wan-v2.6 by Alibaba — AI Model Family

The wan-v2.6 family from Alibaba's Tongyi Wanxiang team represents a breakthrough in AI video generation, delivering ultra-realistic, cinematic videos up to 15 seconds long at 1080p resolution. Unveiled in December 2025, this next-generation series solves key challenges in AI content creation by ensuring scene continuity, character stability, camera rhythm control, and native audio-visual synchronization, enabling professional-grade storytelling without manual editing.

This family encompasses five specialized models: Wan v2.6 Flash (Image to Video) for rapid generation, Wan v2.6 Image to Video for dynamic image animation, Wan v2.6 Text to Video for script-based narratives, Wan v2.6 Image to Image for visual style transfers, and Wan v2.6 Text to Image for high-fidelity static visuals. These models support text-to-video (T2V), image-to-video (I2V), reference-to-video (R2V), image-to-image (I2I), and text-to-image (T2I), making wan-v2.6 a comprehensive toolkit for creators seeking cinematic quality with advanced control.

wan-v2.6 Capabilities and Use Cases

The wan-v2.6 family excels across multiple generation modes, powering everything from quick prototypes to multi-shot narratives with smooth transitions, natural camera movements, and lip-synced audio.

• Wan v2.6 Flash (Image to Video): A distilled version for fast inference, generating 5-15 second videos from images in seconds while preserving motion and audio sync. Ideal for rapid iterations in advertising or social media previews.
• Wan v2.6 Image to Video (I2V): Animates static images into dynamic 5-15 second clips at 720p or 1080p, maintaining subject features, style, and adding realistic motion. Use it to bring product photos to life, like turning a portrait into a talking head video.
• Wan v2.6 Text to Video (T2V): Creates cinematic videos from natural language prompts, supporting multi-shot sequences with cuts, angle changes, and pacing. Perfect for storyboards or ads; example prompt: "A cozy café at dawn: wide shot of barista brewing coffee, cut to close-up of steaming cup, then pan to customer smiling as they sip."
• Wan v2.6 Image to Image (I2I): Transforms input images with precise style control, logical reasoning, and realistic textures for professional edits like relighting or stylization.
• Wan v2.6 Text to Image (T2I): Generates high-definition stills as foundations for video pipelines, with superior artistic control.

These models integrate seamlessly for pipelines: Start with T2I or I2I for keyframe design, feed into I2V or T2V for animation, and refine with R2V using up to 3 reference videos (2-30 seconds) for character consistency, voice cloning, and multi-character scenes. All support 24 fps for widescreen cinematic output, MP4/MOV formats, and durations of 5s, 10s, or 15s (R2V up to 10s).

What Makes wan-v2.6 Stand Out

wan-v2.6 sets new benchmarks with its rebuilt narrative engine, enabling automatic multi-shot storytelling that interprets storyboard prompts into coherent videos with panoramic, close-up, and tracking shots—eliminating post-production stitching. Pioneering China's first Reference-to-Video (R2V) mode, it extracts appearance, motion, voice, and style from reference videos, delivering photorealistic digital avatars with enhanced lip-sync and audio soundscapes for flawless consistency across scenes.

Key strengths include extended 15-second durations (50% longer than predecessors), native AV sync for vocals and effects, 1080p cinematic quality with realistic textures and lighting, and improved instruction-following for complex narratives. Compared to prior versions, it adds video references, multi-shot capabilities, and voice cloning, outperforming in character stability and rhythm control. This makes it ideal for filmmakers, marketers, advertisers, and animators needing high-motion, professional results with minimal effort—especially for character-driven stories, ads, or social content.

Access wan-v2.6 Models via each::labs API

each::labs is the premier platform for seamless access to the full wan-v2.6 family through a unified API, empowering developers and creators to integrate these Alibaba models effortlessly. Run Wan v2.6 Flash for speed, chain T2V with R2V for custom avatars, or scale I2V pipelines—all from one endpoint at eachlabs.ai.

Explore in the interactive Playground for instant testing with sample prompts, or deploy via SDK for production apps. Sign up to explore the full wan-v2.6 model family on each::labs.