Wan 3.0 Image-to-Video
Wan 3.0 Image-to-Video is Alibaba's next-generation AI image-to-video model under the Tongyi Wanxiang series. Supporting single first-frame driving and smooth first-to-last frame transitions, it directly generates up to 30-second videos in up to 1080P/4K cinematic resolution.
Related Models
Wan 3.0 Prime Reference-to-Video Spicy
Wan 3.0 Prime Reference-to-Video Spicy is Alibaba's flagship uncensored high-motion video model engineered for native 4K visual synthesis. Delivering native 4K output at 60 fps, it processes up to 8 reference images or 2 reference videos simultaneously to preserve character identity and material fidelity across complex sequences. With unrestricted generation capabilities, it synthesizes explosive physical movements, dramatic camera trajectories, and physical interactions—powering high-energy cinematic pre-visualization, high-impact commercial ads, AAA gaming assets, and action pipelines.
Wan 3.0 Reference-to-Video Spicy
Wan 3.0 Reference-to-Video Spicy is Alibaba's high-motion video synthesis model engineered for high-energy visual dynamics and stylized action generation. Built on an expanded spatio-temporal attention mechanism alongside robust reference feature anchoring, it synthesizes large-scale physical movements, aggressive camera trajectories, and dramatic temporal transitions while preserving strict subject identity and garment texture fidelity. Wan 3.0 Reference-to-Video Spicy powers dynamic action video production, cinematic FX pre-visualization, interactive gaming visual assets, and high-impact commercial advertising pipelines.
Wan 3.0 Prime Reference-to-Video
Wan 3.0 Prime Reference-to-Video is Alibaba's flagship controllable video generation model built for precise reference-conditioned visual synthesis. Powered by an upgraded spatio-temporal decoupling architecture and multi-reference feature fusion, it preserves character identity, garment textures, ambient lighting, and complex camera trajectories across extended sequences while generating native 4K high-frame-rate video. Operating with strict temporal consistency and fluid motion dynamics, Wan 3.0 Prime drives e-commerce video production, cinematic pre-visualization, digital human animation, and commercial advertising pipelines.
Wan 3.0 Reference-to-Video
Wan 3.0 Reference-to-Video is Alibaba's flagship high-efficiency video generation model engineered for precise reference-driven visual synthesis. Built on an upgraded multi-modal reference feature alignment and spatio-temporal attention architecture, it preserves strict subject identity, garment textures, and ambient lighting across dynamic video sequences. Combining high-throughput inference speeds with fluid camera movement, Wan 3.0 Reference-to-Video efficiently powers e-commerce product videos, digital human animation, social media marketing assets, and commercial advertising pipelines.
Read Me
Wan 3.0 Image-to-Video API
Wan 3.0 Image-to-Video (aliyun/wan3-0/image-to-video) is a state-of-the-art video generation model developed by Alibaba Cloud (Tongyi Wanxiang). Released on August 24, 2026 (with public beta starting August 6, 2026), the model excels in transforming static images into high-fidelity dynamic videos with native audio synthesis. It is accessible via the iCreat asynchronous task endpoint at https://api.icreat.ai/v1/task/submit using Bearer API Key authentication.
The model supports resolutions of 480P, 720P, and 1080P at 30fps, with durations ranging from 2 to 30 seconds. Pricing is based on resolution and duration, starting at $0.05 per second for 480P, $0.10 per second for 720P, and $0.20 per second for 1080P.
Model Positioning
Wan 3.0 Image-to-Video is positioned as a premium, high-resolution multimodal generation model designed for professional creators, marketing agencies, and enterprise applications. It bridges the gap between static imagery and cinematic video content, offering extended generation times (up to 30 seconds) and multi-reference inputs. With the introduction of the Spicy variant, it also caters to users requiring uncensored creative freedom, maintaining the original generative capabilities while relaxing standard content limitations.
Core Capabilities
High-Fidelity Image-to-Video Synthesis The model transforms a single reference image into a fluid, realistic video, maintaining high detail and consistency from the original visual input. It operates at 30 frames per second across 480P, 720P, and 1080P resolutions.
Native Audio Generation Unlike its predecessors, Wan 3.0 natively integrates audio synthesis into the video generation pipeline, eliminating the need for post-generation audio matching and ensuring perfect audio-visual synchronization.
Extended Duration and Multi-Reference Inputs Supporting video durations up to 30 seconds, the model allows users to input up to 10 reference images, 5 reference videos, and 5 reference audio files to guide the generation process, providing unprecedented control over the output.
Flexible Aspect Ratios The model supports various aspect ratios including 1:1, 9:16, 16:9, 3:4, and 4:3, catering to diverse formats from vertical short videos to widescreen cinematic outputs.
Pricing
| Resolution | Unit price (USD/s) | 5-second cost |
|---|---|---|
| 480P | $0.05/s | $0.25 |
| 720P | $0.10/s | $0.50 |
| 1080P | $0.20/s | $1.00 |
Note: Pricing is calculated based on the selected resolution and the actual generated duration. The Prime tier is available at 1.4x the base price. All prices are based on the iCreat platform pricing.
Application Scenarios
- Digital marketing and advertising campaigns requiring dynamic visual content from existing product images.
- Short video creation for social media platforms (TikTok, Instagram Reels, YouTube Shorts) using 9:16 or 16:9 ratios.
- Film and animation pre-visualization, allowing directors to test scenes from storyboards.
- E-commerce product demonstrations, bringing static catalog images to life.
- Creative arts and entertainment, generating extended narrative clips with synchronized native audio.
- Enterprise training and presentation materials requiring high-definition video outputs.
Model Comparison
Comparison Table 1: Same Series (Standard vs Spicy)
| Feature | Wan 3.0 Image-to-Video (Standard) | Wan 3.0 Image-to-Video (Spicy) |
|---|---|---|
| Model Variant | Standard Censored Version | Uncensored Improved Variant |
| Content Policy | Strict content filtering | Relaxed content limitations |
| Core Capabilities | High-fidelity video generation | High-fidelity video generation |
| Native Audio | Supported | Supported |
| Max Duration | 30 seconds | 30 seconds |
| Target Audience | General public, enterprise | Creators requiring unrestricted content |
Comparison Table 2: Cross-Model (Wan 2.7 vs Wan 3.0)
| Feature | Wan 2.7 Image-to-Video | Wan 3.0 Image-to-Video |
|---|---|---|
| Maximum Duration | 15 seconds (2–15s range) | 30 seconds (2–30s range) |
| Supported Resolutions | 720P, 1080P | 480P, 720P, 1080P |
| Audio Support | No native audio | Native audio generation |
| Reference Inputs | Supports first/last frame | Up to 10 images / 5 videos / 5 audio |
| Negative Prompt | Supported | Supported |
| Pricing (720P) | $0.10/s | $0.10/s |
| Pricing (1080P) | $0.15/s | $0.20/s |
| Prime Tier | Not applicable | 1.4x base price |
Why Choose Wan 3.0 Image-to-Video?
- Extended generation time of up to 30 seconds, doubling the capacity of the previous generation.
- Native audio synthesis saves time and resources by generating synchronized audio directly within the model.
- Broader resolution support, introducing 480P for rapid prototyping and cost-effective iterations.
- Comprehensive multi-reference inputs, allowing complex scene composition using images, videos, and audio references.
- Flexible aspect ratio support for various digital platforms and media formats.
- High frame rate of 30fps ensures smooth, cinematic motion quality.
Specifications
| Specification | Details |
|---|---|
| Model Name | Wan 3.0 Image-to-Video |
| Developer | Alibaba Cloud (Tongyi Wanxiang) |
| Endpoint | https://api.icreat.ai/v1/task/submit |
| Release Date | 2026-08-24 (Public Beta: 2026-08-06) |
| Model Type | Asynchronous Video Generation |
| Authentication | Bearer API Key |
| Input: prompt | Required, string |
| Input: negative_prompt | Optional, string |
| Input: media | Optional, array (type: reference_image, url: string) |
| Parameters: resolution | 480P / 720P / 1080P (Optional, default: 720P) |
| Parameters: ratio | 1:1 / 9:16 / 16:9 / 3:4 / 4:3 (Required) |
| Parameters: duration | 2–30 seconds (Required) |
| Parameters: watermark | true / false (Optional, default: false) |
| Output | 30fps video with native audio |
| Billing Unit | Per second based on resolution |
| API Mode | Three-step flow: submit → query-status → get-result |
Architecture
Wan 3.0 Image-to-Video operates on an asynchronous task-based architecture designed to handle high-computational video generation. The workflow consists of three distinct steps: First, the client submits a generation request via POST to the endpoint, including the prompt, media references, and parameters. The API immediately returns a task ID. Second, the client polls the task status using the query-status endpoint to monitor progress (pending, processing, completed, failed). Finally, upon completion, the client retrieves the generated video and audio assets via the get-result endpoint. The model processes the reference images and textual prompts through a multimodal diffusion architecture, outputting high-definition 30fps video with synchronized native audio.
Notes
- The model supports up to 10 reference images, 5 reference videos, and 5 reference audio files simultaneously.
- The Spicy mode is an uncensored version; generated content is not subject to standard content limitations. Users must comply with local laws and regulations when using this mode.
- Watermark is disabled by default but can be explicitly enabled via the parameters field.
- Asynchronous API calls require proper handling of the three-step flow; clients should implement appropriate polling intervals to avoid rate limits.
Frequently Asked Questions
What is the difference between Wan 3.0 Standard and Spicy modes?
Spicy (uncensored mode) is an improved variant of the standard model that maintains the original generation capabilities while relaxing content limitations. It is designed for creators who require unrestricted creative freedom, whereas the Standard mode adheres to strict content filtering policies.
How is pricing calculated for generated videos?
Pricing is calculated by multiplying the unit price of the selected resolution by the duration of the generated video in seconds. For example, a 10-second 720P video costs $1.00 ($0.10/s × 10s). The Prime tier is billed at 1.4x the base price.
Does the model generate audio automatically?
Yes, Wan 3.0 natively supports audio synthesis, generating synchronized sound effects and background audio directly from the visual content and prompt without a separate audio generation step. Users may also guide the audio via the reference audio feature; the final video file contains both video and audio tracks.
What is the maximum duration for a single task?
A single generation task supports up to 30 seconds. For longer videos, split the content into multiple tasks and stitch them together in post-production.