Wan AI 2.2 A14B: Text & Image to Video Online
Bottom line
Alibaba's open-weight Wan 2.2 MoE model: 1-5 second clips, 720p, flat 36-credit price regardless of length. Start from a prompt or a still image in the Text to Video tool; VdoBloom shows the exact credit cost before each Wan AI 2.2 A14B run, with no separate Wan AI account.
Wan 2.2 A14B is the compact, architecture-forward member of Alibaba's Wan lineup on VdoBloom. The A14B name refers to its Mixture-of-Experts design, roughly 14 billion parameters active per generation step, which Alibaba introduced with the Wan 2.2 generation and released as open weights to the research community. VdoBloom runs the hosted version, so there is nothing to download or configure -- you get the same model through the dashboard without a GPU.
It covers both text-to-video and image-to-video from one page, generating clips from 1 to 5 seconds in one-second steps, fixed at 720p, in five aspect ratios: 16:9, 9:16, 1:1, 4:3 and 3:4. Every one of those durations costs the same: 36 credits, whether you render 1 second or 5, because VdoBloom bills the job at a flat per-generation rate rather than per second for this model.
That flat pricing changes how you should use it. There is no cost advantage to rendering a 1-second clip over a 5-second one, so pick the duration your edit actually needs. The model's short ceiling suits transitions, loops and quick prompt tests better than a full scene; for anything longer, Wan 2.6, 2.7 or 3.0 on the same dashboard scale up in resolution and length. VdoBloom shows the 36-credit cost before you generate and refunds automatically if a render fails.
What it costs on VdoBloom
Wan 2.2 A14B is priced as a flat 36 credits per generation -- every duration from 1 to 5 seconds maps to the same 720p rate, so there's no per-second math to do. That's simpler than most Wan variants, which bill per second, but it also means a 1-second test costs exactly what a 5-second final render costs. If you're iterating on a prompt, generate at the duration you actually need rather than starting short to save credits.
For comparison, P-Video on the same dashboard starts at 2 credits for a 1-second 720p clip and 8 credits for 5 seconds -- meaningfully cheaper for short-form work, though it's a different model with its own look and settings. A14B's flat price makes more sense once you're set on the exact clip length you want.
Common questions about VRAM, downloads and open weights
Because Wan 2.2's weights are public, a lot of the search traffic around this model is about running it yourself: VRAM requirements, ComfyUI workflows, GGUF quantization, Mac and AMD compatibility. VdoBloom doesn't cover any of that -- it runs the model on its own infrastructure and exposes it as a simple settings panel, so none of the local hardware questions apply to using it here.
What does carry over: since the underlying weights are open, the model you get through VdoBloom is the same Wan 2.2 architecture that's widely used in the open-source community, just without the setup. If you've used it locally before, the settings (duration, aspect ratio) will look familiar; if you haven't, there's nothing to install.
What creators use Wan AI 2.2 A14B for
- Generate 1-2 second motion inserts and whip transitions for edits without paying for length you would trim away.
- Make looping 1:1 or 9:16 animations for profile headers, stickers and animated backgrounds.
- Test a prompt idea cheaply before committing it to a longer, higher-resolution Wan 2.6/2.7/3.0 render.
- Animate a still image into a short 720p clip when you need speed over resolution.
How to use Wan AI 2.2 A14B on VdoBloom
- 1Open the Text to Video tool and select Wan AI 2.2 A14B from the model picker.
- 2Write your prompt — or upload a starting image for image-to-video.
- 3Choose your duration (1–5s), quality (720p), aspect ratio, review the credit cost, and generate.
Frequently asked questions
What does the A14B in the name mean?
It refers to the model's Mixture-of-Experts architecture, with roughly 14 billion parameters active per generation step. Alibaba introduced this design with the Wan 2.2 generation and published it as open weights.
How much does Wan 2.2 A14B cost on VdoBloom?
A flat 36 credits per generation, no matter which duration from 1 to 5 seconds you pick -- 720p is the only resolution offered. The cost is shown before you generate.
Do my 10 free signup credits cover a Wan 2.2 A14B render?
Not on their own. New accounts start with 10 free credits, and a single A14B generation costs 36, so you would need to top up or earn more credits first. P-Video (from 2 credits at 720p) is a cheaper model to test the free balance on.
Is Wan 2.2 open source?
Yes -- Alibaba released the Wan 2.2 generation's weights openly, which is why it's popular in local ComfyUI setups. VdoBloom runs a hosted copy through its own API, so you get the same model with no download, install or GPU required.
Can I run Wan 2.2 A14B locally instead of through VdoBloom?
Since Wan 2.2 is open-weight, technically yes with your own GPU and setup, but that is outside what VdoBloom offers. VdoBloom's value is skipping that entirely: no VRAM requirement, no installation, just credits and a browser.
Does Wan 2.2 A14B generate audio?
No -- it's a silent video model. For native audio in the Wan family on VdoBloom, use Wan 2.6 Flash or Wan 3.0, which both generate sound alongside the picture.
Why is Wan 2.2 A14B limited to 1-5 seconds?
It's built for short-clip precision rather than full scenes -- every length from 1 to 5 seconds is selectable in one-second steps. That range suits transitions, loops and inserts better than a single long take. For scenes up to 30 seconds, Wan 3.0 is the model on the same dashboard built for that.
How does Wan 2.2 A14B compare to Wan 2.6 or 2.7?
A14B is the budget, short-form option: fixed 720p, 1-5 seconds, flat 36-credit price. Wan 2.6 and 2.7 add 1080p, run up to 15 seconds, and price per second (Wan 2.7 text-to-video starts at 32 credits for a 5-second 720p clip). Pick A14B for quick loops and inserts, 2.6/2.7 for a full scene.
How strict is Wan 2.2 A14B's content filter?
VdoBloom rates it HOT, the platform's most permissive content tier -- the whole Wan family carries this badge. Filtering is looser than on STRICT-rated models like Kling or Veo, though platform-wide rules on illegal content still apply, and a rejected prompt is not billed.
Does Wan 2.2 A14B support image-to-video as well as text-to-video?
Yes, both run from the same model page with identical settings -- 1-5 seconds, 720p, five aspect ratios. Attach an image to animate it, or leave it blank to generate from text alone.