Wan AI 2.2 A14B: Text & Image to Video Online
Bottom line
Wan 2.2 A14B, Alibaba's open-weight MoE video model, hosted: text or image to video, 1-5 seconds at 720p, a flat 36 credits per clip. 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 open-weight Wan model, the Mixture-of-Experts release (about 14 billion parameters active per step) that made Wan popular in local ComfyUI workflows. On VdoBloom you run the hosted version: no GPU, no download, no VRAM to budget. One model handles both text-to-video and image-to-video, and it appears on both the Text to Video and Image to Video tabs.
Clips run 1 to 5 seconds, in one-second steps, at 720p. From text you choose one of five aspect ratios (16:9, 9:16, 1:1, 4:3, 3:4); from a photo the clip follows the photo's shape. Every length costs the same flat 36 credits, $1.80 at the Lite plan rate, shown before you generate and refunded automatically if the job fails. No subscription is needed: a one-time 300-credit pack is $19.50 and never expires.
The flat price cuts both ways. A 1-second clip costs as much as 5 seconds, so always take the length you need. And at 5 seconds, newer Wan models cost less: Wan 3.0 is 32 credits for 5 seconds at 720p with native audio, and Wan 2.6 Flash is 20. A14B earns its place for 1-second inserts, for matching the look of a local Wan 2.2 workflow, and for a silent 720p clip from either text or a photo in one model. Below: specs, costs, the comparison, where it runs, prompting notes and limits.
Wan 2.2 A14B specs at a glance
These values come from the Wan 2.2 A14B endpoint, its Runware service and the credit table. The pixel sizes are the exact dimensions VdoBloom requests for a text-to-video job at 720p.
| Spec | Wan 2.2 A14B on VdoBloom |
|---|---|
| Input | Text prompt, or a photo plus a prompt (a prompt is always required) |
| Clip length | 1, 2, 3, 4 or 5 seconds |
| Resolution | 720p |
| Aspect ratios (text start) | 16:9 (1280x720), 9:16 (720x1280), 1:1 (960x960), 4:3 (1104x832), 3:4 (832x1104) |
| Aspect ratio (photo start) | Follows the uploaded photo |
| Prompt length | Up to 3,000 characters are sent; anything longer is cut |
| Price | 36 credits flat, any length from 1 to 5 seconds |
| Audio | Not listed with audio |
| Content filter | Flexible (HOT label, shared by the Wan family); platform rules still apply |
| Weights | Open (Wan 2.2 generation); VdoBloom runs a hosted copy via Runware |
| Where it runs | Text to Video tab, Image to Video tab, direct photo-effect tabs |
What Wan 2.2 A14B costs: one flat price
Every length maps to the same price row, so there is no per-second arithmetic: 36 credits, or $1.80 at the Lite plan rate of $15 for 300 credits. Bought once, the same 300 credits cost $19.50 and never expire, and they cover eight A14B clips. Plan credits reset each billing cycle. New accounts get 10 free credits, below the 36-credit price.
| Length | Credits | At $0.05 a credit | Effective per second |
|---|---|---|---|
| 1 s | 36 cr | $1.80 | 36 |
| 2 s | 36 cr | $1.80 | 18 |
| 3 s | 36 cr | $1.80 | 12 |
| 4 s | 36 cr | $1.80 | 9 |
| 5 s | 36 cr | $1.80 | 7.2 |
Read more: Credit packs and plans · Credits per second across every video model
Wan 2.2 A14B vs Wan 3.0 vs Wan 2.6 Flash vs Wan 2.7
For short clips, the newer Wan models are priced per second and undercut A14B's flat 36 credits at every length they offer. Wan 3.0 also takes text or a photo and adds native audio; Wan 2.6 Flash needs a photo and adds native audio; Wan 2.7 Text-to-Video starts at 5 seconds. A14B's distinct traits are the 1-second option, the open-weight lineage and silent output.
| Model | Input | Shortest clip | 2 s 720p | 5 s 720p | Audio |
|---|---|---|---|---|---|
| Wan 2.2 A14B | Text or photo | 1 s | 36 cr | 36 cr | Not listed |
| Wan 3.0 Video | Text or photo | 2 s | 13 cr | 32 cr | Native |
| Wan 2.6 Flash | Photo only | 2 s | 8 cr | 20 cr | Native |
| Wan 2.7 Text-to-Video | Text only | 5 s | Not offered | 32 cr | Listed with audio support |
Read more: Wan 3.0 Video · Wan 2.6 Flash · Wan 2.7 Text-to-Video · Wan 2.2 Animate Replace: swap the performer in your footage
Where to run Wan 2.2 A14B on VdoBloom
A14B is one of the few Wan entries that appears on both main tabs, because the same model handles a text start and a photo start.
- Text to Video tab: pick Wan AI, then 2.2 A14B, write the prompt and choose 1 to 5 seconds and the aspect ratio. Resolution is fixed at 720p.
- Image to Video tab: upload a photo and describe the motion; the clip follows the photo's shape, so crop it to the frame you want first.
- Direct photo-effect tabs: effect tabs outside the keyframe group list the full image-to-video roster, A14B included, and send the effect's prompt with your photo.
Read more: What Wan is: the model family explained · Wan AI review (2026)
If you know Wan 2.2 from ComfyUI
Much of the search interest in this model is about running it yourself: VRAM, GGUF quantisations, ComfyUI graphs, Mac and AMD support. None of that applies here; VdoBloom runs the model on its provider's GPUs and exposes a settings panel. What carries over is the model: this is the Wan 2.2 A14B architecture, so prompts you wrote for a local workflow are a reasonable starting point.
What differs: the hosted composer offers 720p only, 1 to 5 seconds, and five aspect ratios for a text start; you do not set steps, samplers, LoRAs or seeds. Prompts are cut at 3,000 characters. If you need those controls, local is the right tool; if you need a clip without a GPU, this is the shortcut.
Prompting Wan 2.2 A14B
Five seconds is one beat, so describe one action and one camera move: subject first, then what it does, then the setting, then the camera. For a loop, ask for motion that ends where it began ("the lantern sways left and back to centre"). For a photo start, describe only the motion, since the photo already fixes the subject and framing. Because every length costs 36 credits, there is no saving in drafting short: generate at the length the edit needs.
What one 300-credit pack buys in short Wan clips
If short clips are the job, count clips per pack rather than credits per clip. A one-time 300-credit pack is $19.50 and never expires; the table divides it by each model's price for its shortest and its 5-second clip at 720p. A14B's count does not change with length because its price does not.
- Pick A14B for a 1-second clip, a silent clip, or a hosted run of the open-weight Wan 2.2 model.
- Pick Wan 3.0 for 2 to 5 second clips from text or a photo at a lower price, with sound.
- Pick Wan 2.6 Flash when you have a photo and want the most clips per pack.
- Pick Wan 2.7 Text-to-Video when the clip should be 1080p or longer than 5 seconds from text.
| Model | Shortest clip | Clips per 300 credits (shortest) | Clips per 300 credits (5 s) |
|---|---|---|---|
| Wan 2.2 A14B | 1 s, 36 cr | 8 | 8 |
| Wan 3.0 Video | 2 s, 13 cr | 23 | 9 |
| Wan 2.6 Flash (photo only) | 2 s, 8 cr | 37 | 15 |
| Wan 2.7 Text-to-Video | 5 s, 32 cr | 9 | 9 |
Limits and honest caveats
On price per clip, A14B is not the economical Wan option: Wan 3.0 costs 32 credits for 5 seconds at 720p with native audio, and Wan 2.6 Flash costs 20 from a photo. The flat price makes short clips expensive per second (36 credits for 1 second). It is 720p only, 5 seconds at most and not listed with audio. A photo start ignores the aspect-ratio choice and follows the photo. Prompts beyond 3,000 characters are cut without warning. The free 10 signup credits do not cover the 36-credit price. Failed jobs are refunded automatically, and a prompt stopped by the content check is stopped before any credits are taken.
VdoBloom does not rank models on subjective quality; whether A14B's look suits the shot is something one 36-credit clip answers.
What creators use Wan AI 2.2 A14B for
- 1-second motion inserts and whip transitions; Wan 3.0 and Wan 2.6 Flash start at 2 seconds and the other Wan text and image-to-video models at 5.
- Testing a prompt on the same Wan 2.2 A14B architecture you run locally, without tying up your own GPU.
- Silent 720p loops in 1:1 or 9:16 for profile headers, stickers and animated backgrounds.
- Animating a still photo into a short clip on a direct photo-effect tab with A14B as the model.
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
How much does Wan 2.2 A14B cost on VdoBloom?
A flat 36 credits per clip, whatever length you choose from 1 to 5 seconds, at 720p. That is $1.80 at the Lite rate of $0.05 a credit. The cost is shown before you generate and refunded if the job fails.
What does A14B mean?
It refers to the model's Mixture-of-Experts design, with roughly 14 billion parameters active per generation step. Alibaba introduced it with the Wan 2.2 generation.
Is Wan 2.2 open source?
Yes. Alibaba released the Wan 2.2 weights openly, which is why it is common in local ComfyUI setups. VdoBloom runs a hosted copy, so you get the model with no download, install or GPU.
Does Wan 2.2 A14B do both text-to-video and image-to-video?
Yes, in one model. It appears on the Text to Video tab (with five aspect ratios) and the Image to Video tab (where the clip follows your photo's shape). Both cost 36 credits for 1 to 5 seconds at 720p.
Why is it limited to 1-5 seconds and 720p?
Those are the options VdoBloom offers and prices for this model: one-second steps from 1 to 5, at 720p. For longer clips, Wan 3.0 runs to 30 seconds; for 1080p, Wan 2.7 or Wan 3.0.
Is Wan 2.2 A14B cheaper than Wan 3.0?
No. Wan 3.0 costs 32 credits for 5 seconds at 720p and 13 for 2 seconds, both below A14B's flat 36, and it adds native audio and longer lengths. A14B's advantages are the 1-second option and matching a local Wan 2.2 workflow.
Does Wan 2.2 A14B generate audio?
It is not listed with audio. For sound generated with the picture, Wan 3.0 (text or photo) and Wan 2.6 Flash (photo) generate audio natively.
Do the 10 free signup credits cover an A14B clip?
No; a clip is 36 credits. Wan 2.6 Flash renders 2 seconds at 720p from a photo for 8 credits, inside the free balance. A one-time 300-credit pack is $19.50, never expires, and covers eight A14B clips.
Can I use my ComfyUI settings, LoRAs or seeds?
No. The hosted composer takes a prompt, an optional photo, the length and (for text) the aspect ratio. Sampler, steps, LoRAs and seeds are not exposed, and prompts are cut at 3,000 characters.
How strict is the content filter?
It carries VdoBloom's Flexible label (HOT in the picker), the most permissive tier, shared by the Wan family. Platform rules still apply, and a prompt stopped by the content check is stopped before any credits are taken.
Wan AI 2.2 A14B guides, prices and comparisons
This page is the Wan AI 2.2 A14B hub. Each guide below goes deeper on one question and links back here.
Wan AI review (2026)
The Wan family on VdoBloom, from 2.2 to 3.0.
WAN vs Seedance
The two biggest families compared on specs and credits.
Wan 3.0 guide
The newest Wan: 2 to 30 seconds with native audio.
Cheapest AI video model
Credits per second at 720p for every video model.
Image to video best practices
Getting clean motion from a still photo.