What Wan 3.0 is best at
2 to 30 seconds, your call
Set any whole-second length from 2 to 30. Two seconds is a real option here, which makes Wan 3.0 unusually good for inserts, transitions, and reaction beats that most models pad out to five.
1080p by default, no 4K
Wan 3.0 renders at 480p, 720p, or 1080p and defaults to 1080p. Drop to 480p for cheap drafts and back to 1080p for the finished clip — the prompt and references stay the same.
Twenty reference inputs
Reference-to-video takes up to 10 images, 5 clips, and 5 audio takes in one request. Address them in the prompt as @Image1, @Video1, or @Audio1 and Genfire rewrites them into the form the model expects.
Start and end frames
Image-to-video begins on the frame you give it and can land on a second frame you choose. Aspect ratio can be set to adaptive so the output matches whatever shape your source image is.
Wan 3.0 at a glance
Prompt ideas for Wan 3.0
Built to Wan 3.0’s strengths — steal these to get started.
How to use Wan 3.0 on Genfire
- 1
Describe it
Write a prompt — or upload an image or reference media, and pick Wan 3.0 in the model selector.
- 2
Generate & compare
Render the video, iterate on the prompt, or run the same brief on a sibling model to compare — all from one credit balance.
- 3
Finish & publish
Send the result into Genfire's editor, upscaler, and audio tools, then export platform-ready output.
What Wan 3.0 changes from WAN 2.5
The first thing to know is that Wan 3.0 is a different kind of model from the WAN 2.5 family that came before it. WAN 2.5 on Genfire is the character-animation workhorse — Animate Move drives a still from a reference video, Animate Replace swaps the subject in existing footage. Wan 3.0 drops those modes and becomes a general-purpose generator with a much longer reach: 2 to 30 seconds in a single pass where the earlier models topped out around 5-10, at up to 1080p with sound, and with a reference pool that WAN 2.5 never had. If you need to puppeteer a mascot from a dance clip, stay on WAN 2.5. If you need a finished 20-second shot with ambience and a consistent cast, Wan 3.0 is the one.
The reference pool is where most of the control lives. A reference-to-video request accepts up to 10 images, 5 clips, and 5 audio takes — 20 files at most — and you cite them positionally in the prompt. In the Genfire studio and through the API you write @Image1, @Video1, or @Audio1, and Genfire normalizes those into the bare "Image 1" form Alibaba's endpoint reads. Images hold a face, a product, or a location; clips carry a camera move or a piece of motion to follow; audio takes give a character a voice or a scene its music. Clips and audio are short-form references, not source footage — the pool is there to condition a new shot rather than to edit an existing one. Pair image references with a locked influencer identity and the same person can appear across a dozen clips with the same voice.
The second axis is the tier. Wan 3.0 and Wan 3.0 Prime accept identical inputs — same durations, same resolutions, same reference limits, same audio toggle — and differ only in the render: Prime is faster and higher-fidelity at 1.4x the price. There are no separate speed tiers beyond that, and no 4K; 1080p is the ceiling and also the default, with 480p and 720p available as cheaper steps. Prompts run to 5,000 characters, so long shot lists with explicit timing fit comfortably. Everything here is live wherever you work: the browser studio, the REST API as video.wan_3 and video.wan_3_prime, the MCP server, and the CLI, all against the same limits.
Loved by creators
“My agency saves 40+ hours per week using GENFIRE. ROI was instant. Absolutely essential tool.”
“The consistency in character generation is miles ahead of anything else I've used.”
“Finally, an AI tool that actually understands cinematic lighting. My portfolio looks amazing.”
Wan 3.0 — frequently asked questions
What is Wan 3.0?
Wan 3.0 is Alibaba's video generation model. It renders clips from 2 to 30 seconds long at 480p, 720p, or 1080p, with synchronized audio generated in the same pass as the picture. It works from a text prompt, from a start image with an optional end frame, or from a pool of reference images, clips, and audio takes.
How long can a Wan 3.0 video be?
Anywhere from 2 to 30 seconds, chosen in whole seconds. The 2-second floor is lower than most video models offer, which makes Wan 3.0 a practical choice for short inserts and transitions as well as for 30-second single-take shots.
What resolution does Wan 3.0 support?
Wan 3.0 renders at 480p, 720p, or 1080p, and 1080p is the default. There is no 4K tier. Lower resolutions cost fewer credits per second, so a common pattern is to draft at 480p and re-run the same prompt and references at 1080p for the final.
Does Wan 3.0 generate audio?
Yes. Wan 3.0 generates synchronized audio — ambience, sound effects, and speech — alongside the video in a single pass. Audio is on by default in every mode, and you can switch it off if you want a silent clip.
How many references can Wan 3.0 take?
Reference-to-video accepts up to 10 reference images, 5 reference clips, and 5 audio takes in a single request. You refer to them in the prompt as @Image1, @Video1, and @Audio1 on Genfire, which rewrites them into the positional form the model expects.
What is the difference between Wan 3.0 and Wan 3.0 Prime?
They take identical inputs — the same durations, resolutions, aspect ratios, reference limits, and audio option. Prime is the faster, higher-fidelity render of the same model and costs 1.4x as much as standard Wan 3.0. Use standard Wan 3.0 for cost-efficient 1080p with sound and Prime when the shot has to look its best.
How is Wan 3.0 different from WAN 2.5?
WAN 2.5 on Genfire is built around character animation, with Animate Move and Animate Replace driving a still image from a reference video or swapping the subject in existing footage. Wan 3.0 is a general video generator instead: 2-30 second clips at up to 1080p with native audio, a start and end frame on image-to-video, and a 10-image, 5-clip, 5-audio reference pool. Both remain available on Genfire.
Compare with other models
From the blog
All posts →Explore Genfire
AI Tools
AI Models
Generate with Wan 3.0 today
Wan 3.0 plus 65+ other AI models, one credit balance, and a full editing pipeline — free plan included, no credit card required.
Start Free