How does Wan 2.2 A14B Turbo select a workflow?

No uploaded media selects text-to-video. One image without audio selects image-to-video. One image plus one audio file selects speech-to-video. Audio alone is not a valid speech input.

What are the six Wan 2.2 A14B Turbo prices?

Text and image workflows output a fixed five seconds for 40/60/80 credits at 480p/580p/720p. Speech rates are 12/18/24 credits per output second; Kyeo calculates seconds as `num_frames ÷ frames_per_second` and rounds the total up.

Which image and audio files are accepted?

Image and speech routes use one JPEG, PNG, or WebP up to 10 MB. Speech also takes one MP3/MPEG, WAV, OGG, M4A, FLAC, AAC, or WMA file up to 10 MB.

Why does 580p appear in all three workflows?

The current text, image and speech workflows all support 480p, 580p and 720p. Their billing differs: text and image are fixed five-second tiers, while speech is billed per output second.

Does enabling the safety check guarantee safe output?

No. Automated checks may miss cases, so every result still requires manual safety and rights review.

How should I compare Wan 2.2 A14B Turbo with Wan 2.6?

Compare the input contracts first: Wan 2.2 has an image-plus-audio speech workflow and 12 workflow-specific controls; Wan 2.6 uses text, image, or video with only resolution and duration controls.

AI video model

Alibaba

Wan

Wan 2.2 A14B Turbo

Wan 2.2 A14B Turbo is not a single Kyeo workflow. No media selects text-to-video, one image selects image-to-video, and one image plus one audio file selects speech-to-video. Each workflow has different resolutions, credit rates, and visible controls, so confirm the media combination first.

Workflows

Text, image, and image-plus-audio

Credit cost

40/60/80 fixed standard tiers; speech 12/18/24 per second

Prompt limit

Prompt required; up to 5,000 characters

Upload limit

Up to 1 image; speech adds 1 audio file; 10 MB per file

Three model versions are selected dynamically by no media, one image, or one image plus one audio file

Text, image and speech workflows all support 480p, 580p and 720p

Text/image fixed five-second tiers are 40/60/80 credits; speech rates are 12/18/24 credits per output second

The speech workflow sends the current `enable_safety_checker` field with a true default

Use this model View credit pricing Back to Model Hub

30-second overview

Wan

Standard 40/60/80; speech 12/18/24 per second

What it does best

Selects text-to-video, image-to-video, or speech-driven video from the supplied media. Each mode uses its own inputs and controls.

Best for

Use it to evaluate prompt-only generation, single-image animation, or image-plus-audio motion as separate tasks. Results from one route should not be generalized to the other two.

At a glance

What is this model like?

Text and image workflows output a fixed five seconds: 40/60/80 credits at 480p/580p/720p. Speech output seconds equal `num_frames ÷ frames_per_second`, multiplied by 12/18/24 credits for the selected resolution; Kyeo rounds the request total up. Speech requires both image and audio, exposes frame, negative-prompt and inference controls, and sends the current `enable_safety_checker` field with a true default. The page does not promise generation speed, synchronization, image quality, or identity consistency.

What it does best

Selects text-to-video, image-to-video, or speech-driven video from the supplied media. Each mode uses its own inputs and controls.

Best for

Use it to evaluate prompt-only generation, single-image animation, or image-plus-audio motion as separate tasks. Results from one route should not be generalized to the other two.

Why use it on Kyeo AI

Kyeo selects one of the three workflows from the supplied media and shows the standard fixed five-second tiers and speech per-second rates before submission; each route still requires separate result validation.

Key facts

Quickly assess whether this model fits your use case.

Also known as

The same model may appear under different names across documentation and community discussions; this list keeps them easy to search and compare.

Wan — Which Wan 2.2 workflow exposes a safety check?

Wan — What happens if Wan 2.2 receives audio without an image?

Wan — How should I compare Wan 2.2 with Wan 2.6?

Common questions

These practical questions focus on the task, input conditions, and result requirements you should confirm before choosing.

How do I use Wan 2.2 A14B Turbo?

Which media select each Wan 2.2 workflow?

How are the six Wan 2.2 credit prices calculated?

Do all three Wan 2.2 workflows support 580p?

Which files does Wan 2.2 image-to-video accept?

Does Wan 2.2 speech-to-video require an image and audio?

Why must Wan 2.2 num_frames be divisible by four?

Which Wan 2.2 workflow exposes a safety check?

What happens if Wan 2.2 receives audio without an image?

How should I compare Wan 2.2 with Wan 2.6?

Selection guide

Use these decision points when choosing a model.

Choose the workflow from the media first

Use no media for text, one image without audio for image-to-video, or one image plus one audio file for speech.

Check resolution and price next

Text and image produce a fixed five seconds for 40/60/80 credits at 480p/580p/720p. Speech costs 12/18/24 credits per output second, calculated from `num_frames ÷ frames_per_second`, and Kyeo rounds the total up.

Validate advanced speech controls

For speech mode, set `num_frames` to a multiple of four from 40 to 120 and frame rate from 4 to 60.

Popular comparisons

Compare common alternatives on the same task to clarify differences in inputs, controls, and cost.

How do Wan 2.2 and Wan 2.6 media workflows differ?

Wan 2.2 vs Wan 2.6: resolution and credits

How do Wan 2.2 and Wan 2.5 Video controls differ?

Wan 2.2 vs Wan 2.5 Video: inputs and pricing

How does Wan 2.2 speech mode differ from Infinitalk?

Wan 2.2 vs Infinitalk: image, audio, and prompt contract

Model comparison

Compare the current model with alternatives at a glance.

Media-routed input

Wan 2.2 A14B Turbo

No media, one image, or one image plus one audio file

Wan 2.6

No media, one image, or one video

Infinitalk

Always requires one image, one audio file, and a prompt

Current credits

Wan 2.2 A14B Turbo

Text/image fixed five seconds: 40/60/80; speech: 12/18/24 per output second

Wan 2.6

70-315 by resolution and duration

Wan 2.5 Video

60-200 by resolution and duration

Resolution options

Wan 2.2 A14B Turbo

Text, image and speech all support 480p, 580p and 720p

Wan 2.6

720p or 1080p

Wan 2.5 Video

720p or 1080p

Visible controls

Wan 2.2 A14B Turbo

Text/image use standard controls; speech adds frame, inference and safety controls

Wan 2.6

Resolution and duration only

Infinitalk

Resolution and seed only

Current validation boundary

Wan 2.2 A14B Turbo

Audio needs an image; frame count must divide by four

Wan 2.6

Its video workflow does not use an image or audio file

Infinitalk

Image and audio are always required together

Practical usage insights

Practical guidance based on public sources, current on-site limits, and representative tasks.

current public source workflow names are not Alibaba direct model IDs

Alibaba's official Wan 2.2 A14B release covers text and image models. Alibaba Cloud's current `wan2.2-s2v` also has a separate image-detection step. The current public source A14B speech workflow must be described only from its current public source form and Kyeo payload.

Marketing claims do not guarantee a specific result

720p, 24 fps, fast generation, stable motion, and audiovisual synchronization are not Kyeo result guarantees; use the actual task output.

Capabilities

Text and image modes

Text mode offers prompt, resolution, aspect ratio, prompt expansion, acceleration, and an optional seed. Image mode accepts one reference image instead of a separate aspect-ratio setting and keeps the other visible controls. Safety checking remains enabled.

Speech-driven mode

Speech mode combines a prompt, one image, and one audio clip, with controls for frame count, frame rate, resolution, inference steps, guidance strength, and shift. Negative prompt and seed are optional, and safety checking is enabled by default.

Safety-field boundary

Safety checking is enabled across all modes, but automated checks do not replace manual content-safety and media-rights review.

Use cases

Prompt-only shot checks

Compare 16:9 with 9:16 and 480p/580p/720p, then review subjects, motion, text, and continuity manually.

Single-image animation

Upload one JPEG, PNG, or WebP and describe the desired action and camera motion.

Image-plus-audio motion

Use one supported image and audio file to test speech or body movement. Before publication, review sync, identity, voice, and rights to every input and output.

Prompt tips

Describe visible action first

State the subject, setting, action, and camera movement without phrasing speed, stability, or final quality as guaranteed outcomes.

Do not assume an aspect ratio in image mode

Image mode does not send aspect_ratio. Inspect the source composition, describe the desired motion, and review the result for cropping and edge artifacts.

Separate audio from visual direction

Audio supplies the sound input; the prompt still describes the person's state and visual action. Negative prompt, guidance, and shift are request controls, not synchronization or quality guarantees.

Why choose it

Media presence deterministically selects one of three workflow IDs.

Text and image expose a compact set of aspect, prompt-expansion, seed, and acceleration controls.

Safety checking is enabled by default in every workflow, including speech

All six workflow-specific Kyeo prices and resolutions are visible before submission.

What to know first

Generate one representative sample with the selected settings and confirm that it meets the goal before expanding the batch.

All three workflows support 480p, 580p and 720p, but standard routes use fixed five-second tiers while speech is billed from `num_frames ÷ frames_per_second`.

Speech-mode `num_frames` must be a multiple of four from 40 to 120.

Turbo speed, 24 fps, motion stability, synchronization, image quality, safety filtering, and completion time vary with the input and current public source state and are not Kyeo result guarantees.

Adjustable parameters

Quickly assess whether this model fits your use case.

Resolution

resolution

Optional

Parameter type: select

Default: 720p

480p

580p

720p

Aspect ratio (text only)

aspect_ratio

Optional

Parameter type: select

Default: 16:9

16:9

9:16

Prompt expansion (text/image)

enable_prompt_expansion

Optional

Parameter type: select

Default: false

Off

Random seed

seed

Optional

Parameter type: number

Range: 0-2147483647

Acceleration (text/image)

acceleration

Optional

Parameter type: select

Default: none

none

regular

Frame count (speech)

num_frames

Optional

Parameter type: number

Default: 80

Range: 40-120; must be divisible by 4

Frames per second (speech)

frames_per_second

Optional

Parameter type: number

Default: 16

Range: 4-60

Negative prompt (speech)

negative_prompt

Optional

Parameter type: text

Inference steps (speech)

num_inference_steps

Optional

Parameter type: number

Default: 27

Range: 2-40

Guidance scale (speech)

guidance_scale

Optional

Parameter type: number

Default: 3.5

Range: 1-10

Shift (speech)

shift

Optional

Parameter type: number

Default: 5

Range: 1-10

Safety check (speech only)

enable_safety_checker

Optional

Parameter type: select

Default: true

Off

Credit usage

Standard 40/60/80; speech 12/18/24 per second

Budget tip

Before batch generation, run an A/B test with the same assets across the current model and alternatives to validate quality and avoid wasting credits.

FAQ

Related models

Compare these similar candidates before deciding.

Wan 2.6

No media selects text-to-video, exactly one image selects image-to-video, and one to three videos select video editing; images and videos cannot be mixed. The default is 1080p, 5 seconds, shot structure Off, and content filtering On.

AI video model

70–315 credits / request

View details Use now

Wan 2.5 Video

Review Wan 2.5 Video on Kyeo AI with text-to-video or one-image input, 5/10 seconds, 720p/1080p, prompt constraints, and 60–200-credit pricing. Kyeo charges by resolution and duration: 60 credits for 720p at 5 seconds, 120 for 720p at 10 seconds, 100 for 1080p at 5 seconds, and 200 for 1080p at 10 seconds. The 720p 5-second default costs 60 credits. How are Wan 2.5's 60/100/120/200 credits calculated? — Review Wan 2.5 Video on Kyeo AI with text-to-video or one-image input, 5/10 seconds, 720p/1080p, prompt constraints, and 60–200-credit pricing.

AI video model

60/100/120/200 credits per request

View details Use now

Infinitalk

Infinitalk uses 1 image, 1 audio file no longer than 15 seconds, and a required prompt to generate a speaking-portrait video on Kyeo. It offers 480p or 720p plus optional `seed`; cost depends on audio duration and resolution rather than a fixed request price.

AI video model

480p: 3 credits/sec; 720p: 12 credits/sec

View details Use now

Sources

Content is based on model documentation, feature references, and the settings available on Kyeo AI.

Source note

Review Wan 2.2 A14B Turbo text, one-image and speech-driven workflows, fixed five-second tiers, per-second speech rates and mode-specific inputs. Text and image workflows output a fixed five seconds: 40/60/80 credits at 480p/580p/720p. Speech output seconds equal `num_frames ÷ frames_per_second`, multiplied by 12/18/24 credits for the selected resolution; Kyeo rounds the request total up.

Last updated: 2026-08-27

Alibaba Cloud: Wan2.2 A14B release

Official

View source

Alibaba Cloud: direct wan2.2-s2v guide

Official

View source