ModelFoldwan/3-0-video
Wan3.0 is Alibaba’s next-generation all-in-one video generation model built for multimodal creation. It supports text, image, video, audio, and structured references to produce more consistent, realistic, and controllable video content with enhanced storytelling and editing capabilities.
Pricing: From 10 credits / sec at 480p (50 credits for 5 seconds), billed on output seconds. Every resolution is in the rate card below.
Explore more Text to Video models
More AI models from Wan
Latest and popular AI models
Wan3.0 API Redefines Multimodal Video Creation with Greater Control
Experience the next evolution of video generation with Wan3.0 API, combining longer narratives, stronger reference consistency, precise editing, and immersive audiovisual creation for more complete visual storytelling.
Start Video Creation from Richer Multimodal Inputs with Wan3.0 API
Wan3.0 is Alibaba’s latest all-in-one video generation model, designed for longer-form storytelling, richer reference understanding, stronger visual consistency, and more precise creative control. Wan3.0 API enables creative generation based on different types of input materials, including text, images, videos, audio, keyframes, documents, and webpages, helping shape scenes, motion, sound, visual identity, and narrative direction through more flexible references. For the latest supported inputs, generation modes, and API usage details, please refer to the Wan3.0-Video API documentation on ModelFold.
Turn Text Prompts into Full Video Concepts
Text can do far more than describe a simple scene. With Wan 3.0 API, written instructions can shape characters, environments, camera movement, action logic, lighting, pacing, atmosphere, and even the emotional rhythm of a video, making detailed prompts a flexible starting point for more structured creation. Stay with Wan 3.0 Video when its multimodal video workflow is enough, and move to the newer MiniMax H3 (Hailuo-03) API when native stereo sound alongside video generation is the deciding factor.
Carry Visual Identity into Motion with Image References
Image references help define the look of a video before motion begins. Character appearance, clothing, props, locations, materials, composition, and overall style can all be guided visually, allowing Wan3.0-Video API to bring recognizable details into moving scenes with stronger continuity.
Use Video References to Guide Movement and Scene Flow
Some creative intent is easier to show than to explain. Existing footage can provide motion language, performance cues, timing, camera behavior, and scene dynamics that help Wan3.0 API build or transform videos with more controlled movement and a clearer sense of flow.
Connect Sound Direction with Visual Generation
Audio can shape the feel of a scene just as much as visuals do. Speech, music, rhythm, vocals, and other sound references give Wan 3.0 Video API more context for pacing, mood, performance energy, and audiovisual coordination, helping the final result feel more unified.
Keep Scene Progression Aligned with Keyframes
Keyframes provide strong visual anchors for how a sequence should begin, evolve, or conclude. By using opening or ending visual references, Wan3.0 API can guide subject continuity, transition logic, and scene development in a way that keeps the generated video more directionally consistent.
Bring Documents and Webpages into the Creative Workflow
Creative source material does not always start as a prompt. Briefs, presentations, spreadsheets, PDFs, text documents, and webpages can also provide valuable structure and context, allowing Wan 3.0 API to turn existing information into part of the video generation workflow instead of treating it as separate preparation work.
Discover What Makes Wan3.0 API a Major Leap in Video Creation
All-in-One Multimodal Creation with Wan3.0 API
Wan3.0 expands video creation beyond conventional prompts by incorporating richer creative context into the generation process. Different types of reference materials, including images, videos, audio, documents, and webpages, can help guide scene development, visual consistency, and creative direction. For the latest supported input types, available features, and API usage details, please refer to the Wan3.0-Video API documentation on ModelFold.
Native 30-Second Storytelling with Wan 3.0 API
Longer native generation gives scenes more space to establish context, develop actions, transition between story beats, and reach a more satisfying conclusion. Wan 3.0 API supports videos up to 30 seconds while intelligently adapting duration to the requested content, opening more room for narrative structure and creative freedom.
Pixel-Perfect Reference Consistency in Wan3.0-Video API
Characters, clothing, props, environments, audio cues, and spatial relationships can remain closely aligned with their original references throughout changing actions and scenes. Wan3.0-Video API strengthens detail reproduction and continuity, making reference-driven creation more dependable when visual identity needs to remain stable.
Precision Video Editing with Wan 3.0 Video API
Existing videos can be reshaped through detailed creative instructions without unnecessarily changing everything in the original footage. Wan 3.0 Video API supports precise adjustments to subjects, backgrounds, environments, lighting, visual styles, and other elements, alongside reference-guided editing and video extension for more flexible creative refinement.
Immersive Audiovisual Generation through Wan3.0 API
Visual storytelling is paired with more expressive audio generation to make scenes feel more complete. Wan3.0 API improves natural timbre, regional speech expression, musical rhythm, environmental sound, and audiovisual synchronization, allowing motion, performance, atmosphere, and sound to develop with stronger cohesion.
Cinematic Visual and Motion Control with Wan3.0-Video API
Wan3.0 raises the level of realism and aesthetic control across lighting, color, composition, character performance, camera language, and scene dynamics. Wan3.0-Video API can handle complex action choreography, realistic physical interactions, detailed materials, and nuanced performances with greater visual precision for more cinematic results.
How to Start Creating Videos with Wan3.0-Video API on ModelFold
- 1
Set Up Your ModelFold Account for Wan3.0-Video API Access
Create a ModelFold account or log in to your existing account to get started with Wan3.0-Video API. After accessing your API credentials, you can configure your request environment and prepare the necessary settings for connecting video generation capabilities to your application.
- 2
Prepare Creative Inputs for Wan3.0-Video API Requests
Create video generation requests by combining prompts with supported creative references and instructions. Wan3.0-Video API is designed to work with different forms of creative context, helping guide visual consistency, scene development, motion design, and overall storytelling direction. Refer to the latest ModelFold documentation for detailed input options and request formats.
- ✓
Submit Wan3.0-Video API Tasks and Access Generated Results
Send generation requests through Wan3.0-Video API on the ModelFold video endpoint. The response returns the generated video URL once the task finishes, so you can integrate the results into your creative projects or applications without polling.
Exploring the Evolution of AI Video Generation with Wan and Seedance
From cinematic storytelling to more controllable visual creation, Wan 3.0, Seedance 2.5, and Seedance 2.0 represent different approaches to AI video generation. Compare how each model handles motion, consistency, and creative expression through the same visual concept.
Wan 3.0: Next-Level Control for Cinematic Video Creation
Wan 3.0 brings stronger visual understanding and generation control, enabling more consistent characters, environments, and complex scenes. The video showcases improved motion coherence, richer audiovisual details, and more precise creative direction for high-quality video production.
Seedance 2.5: Enhanced Creativity with More Refined Video Dynamics
Seedance 2.5 advances video generation with improved reference understanding, smoother motion patterns, and stronger scene continuity. It offers more expressive camera movement and detailed visual storytelling, bridging the gap between creative exploration and production-ready results.
Seedance 2.0: Reliable Motion and Consistent Visual Storytelling
Seedance 2.0 focuses on stable video generation with strong prompt alignment and smooth scene transitions. It maintains clear subject identity and natural movement, making it a balanced choice for creating engaging visual narratives.
Bring Longer, Richer, and More Controlled Video Ideas to Life with Wan3.0 API
Build Longer Narrative Sequences with Wan3.0 API
A short film can develop across multiple actions, locations, camera changes, and story beats without being reduced to a single isolated moment. Wan3.0 API is well suited to sequences such as an urban chase, miniature adventure, surreal comedy, or cinematic action scene where characters, props, environments, and narrative logic need to remain connected as the story progresses.
Create Reference-Driven Character Videos with Wan 3.0 API
When several characters, costumes, environments, or visual references need to remain recognizable throughout a scene, Wan 3.0 API can use those materials as persistent creative guidance. This makes it useful for fashion films, choreography sequences, character showcases, historical scenes, and other productions where appearance, styling, spatial relationships, and movement continuity all matter.
Rework Existing Footage with Wan3.0-Video API
Existing video can become the starting point for controlled creative changes rather than being regenerated from scratch. Wan3.0-Video API can support tasks such as replacing a subject, changing an environment, removing an accessory, shifting the visual style, adjusting selected scene elements, or extending an existing sequence while preserving the motion and structure that should remain unchanged.
Produce Audio-Visual Content with Wan 3.0 Video API
Videos built around dialogue, music, environmental sound, rhythmic motion, or expressive atmosphere can benefit from treating audio as part of the creative direction from the beginning. Wan 3.0 Video API can support content such as performance videos, cinematic commercials, immersive story scenes, and rhythm-driven motion pieces where sound, visual pacing, character performance, and scene dynamics need to feel closely synchronized.
Why Choose ModelFold for a More Practical Wan3.0 API Workflow
One Model for Text, Frames, References, Files and Links
wan/3-0-video runs on POST https://ai.modelfold.online/v1/videos/generations and covers text-to-video, first-frame and first-and-last-frame video, reference-to-video, and video from a document or a public web page. Output goes from 480P to 1080P and from 2 to 30 seconds with audio on by default; the full parameter list is in the Playground and the API section on this page. Pricing for each resolution is shown in the pricing line above and in the rate card of the API section.
Pick One Input Mode per Request
First and last frames cannot be combined with reference media, a document or a web page, and a last frame needs a first frame. A document and a web page cannot be used together. In reference mode, refer to the media as Image1, Video1 or Audio1 in the prompt, in the order you pass them. Frame and reference images cannot be PNGs with transparency.
Reference Videos Count Toward Duration and Billing
Billed seconds equal the output duration plus the total duration of reference videos, at the same per-second price, and with reference videos the input plus output duration must stay within 30 seconds. With automatic duration (-1), the request is charged for the 30-second maximum. New ModelFold users receive 80 free credits for testing.
Frequently Asked Questions About Wan3.0 API
What is Wan3.0?⌄
Wan3.0 is Alibaba’s all-in-one video generation model designed for longer video creation, multimodal reference understanding, stronger consistency, precision editing, immersive audiovisual generation, and cinematic control. It supports multiple creation approaches, including text-driven, image-driven, reference-based, and existing-video workflows.
Is Wan3.0 API available on ModelFold now?⌄
Yes. Wan3.0 API is now available on ModelFold. Users can access Wan3.0-Video API for video generation and integrate its capabilities into their applications through ModelFold’s API workflow. For the latest supported features and usage details, please refer to the Wan3.0-Video API documentation on ModelFold.
What types of video generation does Wan 3.0 API support?⌄
Wan 3.0 API supports text-to-video, image-to-video, first-and-last-frame video generation, and reference-based video creation. Its broader multimodal capabilities also make it suitable for workflows that combine different creative references within the same generation process.
What kinds of references can Wan3.0-Video API understand?⌄
Wan3.0-Video API supports richer creative context beyond traditional text prompts, allowing different types of references to help guide video generation. Depending on the available input options, reference materials such as images, videos, audio, keyframes, documents, and webpages can contribute to visual consistency, scene structure, motion design, and overall creative direction. For the latest supported reference types and input requirements, please refer to the Wan3.0-Video API documentation on ModelFold.
Can Wan3.0 API generate videos up to 30 seconds?⌄
Yes. Wan3.0 API supports native video generation up to 30 seconds, giving scenes more room for action development, camera progression, character interaction, and complete narrative structure without relying entirely on multiple short clips.
How does Wan 3.0 API maintain reference consistency?⌄
Wan 3.0 API is designed to preserve important details from supplied references across changing scenes and motion. Character appearance, clothing, props, environments, audio characteristics, and spatial relationships can remain more closely aligned as the generated sequence develops.
Can Wan3.0-Video API edit existing videos?⌄
Yes. Wan3.0-Video API supports instruction-based and reference-guided video editing. Existing footage can be modified by changing subjects, environments, visual styles, lighting, and other selected elements while preserving parts of the original video that should remain unchanged. Video extension is also supported for continuing existing sequences.
Does Wan 3.0 Video API support synchronized audio and video creation?⌄
Yes. Wan 3.0 Video API is designed for richer audiovisual generation, including natural sound expression, speech characteristics, music and rhythm, environmental audio, and closer synchronization between sound and visual pacing. This helps generated videos feel more complete and expressive as a unified audiovisual experience.
| 480p | 50 credits per 5 seconds |
| 720p | 100 credits per 5 seconds |
| 1080p | 200 credits per 5 seconds |
Billed per second. The price above is for a 5-second clip at each resolution — longer clips cost proportionally more, and the Playground charges the exact same way.
With reference videos, billed seconds = output seconds + total reference-video seconds at the same per-second price (reference videos up to 15 s in total, each rounded up to a whole second; input + output up to 30 s). Duration -1 (auto) is charged as 30 seconds.
curl https://ai.modelfold.online/v1/videos/generations \
-H "Authorization: Bearer $MODELFOLD_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"wan/3-0-video","prompt":"...","resolution":"720p","duration":5}'