Wan icon

Wan 3.0

3.0

Generate native 30-second videos in Wan3.0 Public Beta for longer explainers, campaign clips, and short narrative scenes Turn documents, spreadsheets, slides, webpages, and other static sources into video with the new Omni-Reference workflow Combine up to 20 reference assets, including images, videos, audio, documents, and webpages, to guide scenes and visual style

Reviewed by ToolWorthy Editors·updated today·3.0 released yesterday

Jump to section
No media available

More tools to compare

InVideo icon

InVideo

AdAnt AI icon

AdAnt AI

Cardboard icon

Cardboard

Kapwing AI icon

Kapwing AI

VEED.IO icon

VEED.IO

Pixlie icon

Pixlie

Pros & Cons

Pros

  • Longer native output with 30-second video generation, reducing the need to combine several short clips manually
  • Richer source inputs through Omni-Reference, including documents, spreadsheets, slides, webpages, and traditional media references
  • Better fit for business content because static reports, product pages, and decks can become direct video inputs
  • Public beta availability gives creators and teams early access to test the next Wan generation
  • Official Alibaba positioning indicates a major model-family update rather than a minor snapshot refresh

Cons

  • Beta-stage uncertainty around quotas, pricing, region availability, API names, and production stability
  • Limited public documentation compared with Wan 2.7 entries in Alibaba Cloud Model Studio documentation
  • No official benchmark table yet for visual quality, speed, prompt adherence, or comparison with previous Wan versions
  • Commercial usage terms need confirmation before using generated videos in regulated or high-volume production workflows
  • Third-party pages may overstate capabilities that Alibaba has not yet documented, so source hygiene is important

Overview

Wan 3.0 is Alibaba's public beta release of its Wan video generation model family, announced on August 6, 2026. The update moves Wan beyond short prompt-to-video generation by adding native 30-second video output and a broader input system called Omni-Reference. Instead of relying only on text prompts, images, audio, or video references, Wan 3.0 can work from document-like and web-based source material, including documents, spreadsheets, slides, and webpages.

This makes Wan 3.0 especially relevant for teams that need to turn dense information into video: product marketers converting landing pages into explainers, educators turning slides into short lessons, or enterprise teams transforming static reports into visual summaries. The release is still in public beta, and Alibaba's detailed Model Studio API, benchmark, and licensing documentation has not fully caught up with the announcement yet.

What's New

Native 30-Second Video Generation

Wan 3.0 introduces native 30-second video generation, extending the timeline for a single output compared with earlier Wan 2.x workflows. Wan 2.6 focused on cloud-based generation with 5, 10, and 15-second text-to-video options, while Wan 3.0 doubles the practical scene length described in Alibaba's announcement.

For creators, the longer native duration matters because it reduces the need to stitch several short generations together. A 30-second window is enough for a short product demo, a social ad, a lesson segment, or a simple narrative sequence with setup and payoff. It should also help teams preserve continuity when a scene depends on consistent subject movement, camera timing, or message pacing.

Omni-Reference Inputs

The biggest workflow change is Omni-Reference. Alibaba describes it as moving beyond text, images, audio, and video references to include documents, spreadsheets, slides, webpages, and other static or text-heavy sources. In practical terms, Wan 3.0 is positioned less like a prompt-only video generator and more like a system that can convert existing business or educational materials into video.

This is a meaningful shift from the Wan 2.x pattern. Earlier versions already supported multiple video generation tasks, including text-to-video, image-to-video, reference-to-video, and video editing through different model endpoints. Wan 3.0's Omni-Reference framing suggests a more unified creation flow where the source material can be richer and more structured before the model generates video.

Reality-Grade Rendering

Alibaba's announcement highlights "reality-grade rendering" as a core quality improvement. The company has not yet published detailed side-by-side benchmarks for Wan 3.0, so this should be read as an official positioning claim rather than a quantified quality score.

The likely user-facing goal is clearer: videos should look more coherent and production-ready when generated from longer prompts and richer reference material. This matters for marketing and training videos because longer clips expose more weaknesses in motion stability, lighting consistency, subject identity, and scene transitions.

Public Beta Release

Wan 3.0 is available as a public beta rather than a fully documented generally available API release. Alibaba announced the beta across its hosted Wan, Qwen Cloud, and Alibaba Cloud surfaces, although availability may vary while access continues rolling out. That matters for production planning because public beta releases can change behavior, quotas, pricing, model names, and access rules while the product team gathers usage feedback.

Teams evaluating Wan 3.0 should treat early outputs as test material until Alibaba publishes stable Model Studio documentation for the exact API names, supported regions, generation limits, result retention, and commercial usage terms. Alibaba Cloud has published public pricing for earlier Wan 2.x video models and continued listing Wan 2.7 entries as the latest documented Wan API generation at the time of this release write-up, so production API users should verify the current console and docs before budgeting Wan 3.0 usage.

Availability & Access

Wan 3.0 was announced through Alibaba Group and the official Wan account on X on August 6, 2026. The announcement describes the model as Public Beta and points users back to the Wan ecosystem rather than a fully detailed API changelog.

The safest access assumptions are:

  • Web access: Use the official Wan website for public beta exploration where available.
  • API access: Check Alibaba Cloud Model Studio before production integration, because full Wan 3.0 API access and documentation may continue rolling out after the social announcement.
  • Status: Treat the release as beta until Alibaba publishes stable API names, limits, pricing, and terms.
  • Source validation: Avoid relying on third-party "Wan 3.0" websites for exact limits or licensing unless Alibaba confirms the same details.

Pricing & Plans

Alibaba has not published complete Wan 3.0 public beta pricing details in the Model Studio lifecycle documentation at the time of writing. Earlier Wan cloud models are generally accessed through Alibaba Cloud Model Studio or the Wan web experience, with usage-based billing depending on model type, duration, resolution, and region. Alibaba Cloud's current pricing pages provide a useful baseline for Wan 2.x models, but those prices should not be treated as confirmed Wan 3.0 rates unless the console or official documentation names the exact Wan 3.0 model.

For now, plan around three practical cases:

  • Public beta exploration: Use the official Wan web experience for hands-on testing if your account has access.
  • Cloud API deployment: Verify Model Studio model names, supported regions, generation limits, and billing units before budgeting production usage, because beta access can change.
  • Enterprise use: Confirm commercial terms directly through Alibaba Cloud if generated videos will be used in paid campaigns, customer-facing education, or automated content pipelines.

Do not assume Wan 3.0 has the same pricing, free quota, open-weight availability, or self-hosting terms as earlier Wan releases until Alibaba publishes those details.

Best For

  • Marketing teams turning landing pages, product sheets, and campaign briefs into 30-second promotional videos
  • Educators and trainers converting slide decks or structured notes into short explainer videos
  • Enterprise content teams summarizing documents, reports, or web pages as internal video briefings
  • Social media creators who need longer clips than earlier short-form Wan outputs without stitching multiple generations
  • Developers and product teams evaluating whether document-to-video generation can fit future automated workflows

FAQ

Is Wan 3.0 officially released?

Yes, Alibaba announced Wan 3.0 on August 6, 2026, and the official Wan account described it as Public Beta. That means the model is real, but some implementation details may still change while Alibaba finalizes documentation and rollout.

What are the main Wan 3.0 features?

The confirmed announcement-level features are native 30-second video generation, reality-grade rendering, and Omni-Reference inputs. Omni-Reference expands source material beyond text, images, audio, and video to documents, spreadsheets, slides, webpages, and other static or text-heavy inputs.

Is Wan 3.0 available through Alibaba Cloud Model Studio?

Alibaba announced Wan 3.0 across its hosted Wan, Qwen Cloud, and Alibaba Cloud surfaces. Alibaba's Model Studio lifecycle documentation had not yet listed a dedicated Wan 3.0 API entry when this page was prepared; the latest documented Wan entries visible there were Wan 2.7 models, so production API users should verify Model Studio before integrating Wan 3.0 into automated systems.

Is Wan 3.0 open source?

Alibaba has not confirmed Wan 3.0 open-weight availability in the official announcement text used for this release page. Earlier Wan releases included open-source components and model weights, but Wan 3.0's exact licensing and self-hosting status should not be assumed until Alibaba publishes a model card, repository, or terms page.

Can I use Wan 3.0 for commercial videos?

Possibly, but confirm the terms first. Public beta access does not automatically mean unrestricted commercial usage. If you plan to use Wan 3.0 outputs in ads, paid courses, client campaigns, or customer-facing product content, check Alibaba Cloud and Wan terms before production use.

Version History

3.0

Current Version

Released on August 6, 2026

+What's new
3 updates
  • Generate native 30-second videos in Wan3.0 Public Beta for longer explainers, campaign clips, and short narrative scenes
  • Turn documents, spreadsheets, slides, webpages, and other static sources into video with the new Omni-Reference workflow
  • Combine up to 20 reference assets, including images, videos, audio, documents, and webpages, to guide scenes and visual style

2.7

Released on April 7, 2026

+What's new
3 updates
  • Direct multi-shot video narratives with stronger camera control, scene continuity, subject consistency, and synchronized audio across complete sequences
  • Generate videos from text, a first frame, first and last frames, or video continuation inputs, supporting more production workflows in one model family
  • Create 720P or 1080P videos up to 15 seconds with improved prompt following, motion quality, audiovisual synchronization, and narrative coherence

2.6

Released on December 16, 2025

View Update
+What's new
3 updates
  • Create character-led videos from reference images or clips while preserving appearance, movement, voice, and background details across new scenes
  • Generate multi-shot narratives up to 15 seconds with stronger subject consistency, cinematic visuals, studio-grade audio, and reliable audiovisual synchronization
  • Use text, images, video, or audio as references to guide characters, props, motion, visual style, and sound within a unified creative workflow

2.5

Released on September 24, 2025

+What's new
3 updates
  • Generate synchronized dialogue, music, ambient sound, and effects with video, producing complete audiovisual clips without adding audio in a separate workflow
  • Create coherent videos up to 10 seconds from text or images with stronger instruction following, motion quality, visual detail, and cinematic output
  • Generate and edit images with multilingual text rendering, multi-image references, improved consistency, and more precise instruction control

2.2

Released on July 28, 2025

+What's new
3 updates
  • Generate videos with cinematic aesthetics using Mixture-of-Experts (MoE) architecture and controllable style tags for lighting, composition, and color tone, enabling advertising teams to maintain consistent visual styles across campaigns
  • Produce 720P videos at 24fps on consumer-grade RTX 4090 GPUs with the new TI2V-5B model and efficient VAE (16×16×4 compression), making high-quality video generation accessible to small studios
  • Access production-ready models through open-source inference code, Diffusers integration, ComfyUI support, and Hugging Face Spaces, allowing developers to choose code-based, GUI, or cloud API workflows based on team needs

2.1

Released on February 25, 2025

+What's new
3 updates
  • Deploy AI video generation privately with released inference code and model weights, running 5-second 480P videos in 4 minutes on RTX 4090 using only 8.19GB VRAM with the T2V-1.3B model
  • Create complete content pipelines from cover images to edited videos using the multi-task suite (T2V, I2V, Video Editing, T2I, Video-to-Audio), eliminating the need for multiple tools
  • Generate videos with embedded Chinese and English text for subtitles, signage, and UI elements, making it the first open video model capable of bilingual visual text rendering for advertisement and product demo content

Top alternatives

Related categories

From the blog

View all →

Track Wan in ToolWorthy Weekly

Important tool updates, better alternatives, and selected AI signals in one weekly brief.

Weekly only. Unsubscribe anytime.