Qwen Create Limits, Formats, Exports & Common Errors

Qwen Create does not have one publicly documented limit table that can be applied to every user, account, image task, and video task. The official creative interface provides separate AI Image, AI Video, and Assets areas, but the public product page does not publish a universal daily quota, guaranteed export format, maximum duration, or asset-retention period for every account.

For developer use, Alibaba Cloud Model Studio publishes model-specific specifications. Current Qwen-Image APIs return PNG images, while current Wan video APIs return MP4 video results with model-specific resolution and duration limits. API image and video result links are generally temporary and must be saved before they expire.

Important distinction: A limit documented for a Model Studio Model ID is not automatically a Qwen Create UI limit. Qwen Create, Qwen Studio, Model Studio APIs, third-party providers, and locally hosted models are separate access routes with different controls.

Qwen Create Limits and Formats at a Glance

Question	Reliable answer
Does Qwen Create have a fixed daily limit?	No single public quota table guarantees the same allowance for every account. Use the counter, notice, or error shown in your current interface.
What image format does the Qwen-Image API return?	PNG under the current official Model Studio documentation.
Can Qwen-Image APIs generate multiple results?	Some Qwen-Image 2.0 and 3.0 models support one to six outputs; other models are fixed at one.
What video format do current Wan APIs return?	Current documented Wan video workflows return MP4 video; several model families specify H.264 encoding and 30 fps.
How long do API result links remain available?	Current Qwen-Image and Wan API documentation gives image URLs, video URLs, and task data a 24-hour validity period.
Does Qwen Create export WebP?	The public Qwen Create page does not publish a universal export-format table. The documented Qwen-Image API output is PNG; convert a downloaded copy to WebP separately when needed.
Why does Qwen video generation fail?	Common causes include a quota or rate limit, invalid duration or ratio, unsupported input, conflicting generation modes, moderation, an expired asset URL, or temporary service capacity.

What Is Qwen Create?

Qwen Create is an official Qwen creative interface with separate areas for AI Image, AI Video, and Assets. It is designed for interactive visual creation rather than normal text chat.

Qwen Create should not be treated as another name for every Qwen image or video service.

Access route	Primary purpose	Where limits come from
Qwen Create	Interactive AI image and video creation with an Assets area	The current product interface, account state, and available creative mode
Qwen Studio	General chat, multimodal understanding, image generation, files, and tools	The Qwen Studio interface and account availability
Alibaba Cloud Model Studio	Developer APIs for Qwen-Image, Wan video models, and other generation services	The exact Model ID, region, endpoint, rate limit, and API documentation
Third-party provider	A separate website or API serving a Qwen or Alibaba model	The provider’s own plan, implementation, model version, and policies
Local model	Generation through downloadable weights and a local runtime	GPU memory, software, model files, workflow settings, and storage

The exact model behind a Qwen Create task may not be exposed as the same Model ID used in Alibaba Cloud Model Studio. Do not copy an API limit into a Qwen Create article unless the product itself confirms that the same limit applies.

Why There Is No Single Universal Qwen Create Limit

The phrase “Qwen Create limit” can refer to several different restrictions:

The number of image or video tasks an account can submit.
The number of tasks that can run simultaneously.
The maximum number of output images.
The maximum image dimensions or total pixel count.
The maximum input-file size.
The maximum number of reference images.
The maximum video duration.
The available output resolution.
The generation queue or rate limit.
The time an output remains available in the interface or through a URL.

These restrictions are not controlled by one number. A user can have enough account quota but still submit an unsupported file, request an invalid duration, trigger content moderation, or encounter a temporary server-capacity error.

How to Check Your Current Qwen Create Limit

Do not rely on an old screenshot, a community post, or another user’s account. Check the value shown in your own Qwen Create interface at the time of the task.

Sign in to the account you will use.
Open the correct AI Image or AI Video area.
Check for a visible counter, credit balance, plan notice, or Limit Reached message.
Record the date, platform, account type, and exact wording of the notice.
Check whether image and video tasks use separate allowances.
Check whether failed tasks change the displayed balance.
Do not publish the observed number as a universal Qwen limit unless Qwen documents it publicly.

If predictable throughput is required for a production application, use a documented API plan rather than designing the workflow around an undocumented consumer-interface allowance.

Qwen Image Input Formats and Limits

The public Qwen Create landing page does not provide a complete file-format table for its logged-in AI Image interface. When an image upload fails, the safest diagnostic method is to create a clean copy in a common format such as PNG, JPEG, or WebP and test one image at a moderate resolution.

Developer-side Model Studio documentation provides more specific requirements. These are useful references, but they should not be presented as guaranteed Qwen Create UI limits.

Documented Qwen Image 3.0 Input Requirements

The current Qwen Image 3.0 API documentation supports one to three reference images for image-to-image or editing requests.

Supported input formats: JPG, JPEG, PNG, BMP, TIFF, WebP, and GIF.
Reference-image count: One to three.
Input file size: Up to 10 MB per image.
Recommended dimensions: Width and height between 384 and 2048 pixels for best results.
Prompt structure: One user message containing the image objects and one text instruction.

An animated GIF should not be assumed to remain animated during image editing. In current Qwen-Image editing documentation, only the first frame of an animated GIF is processed.

Recommended Diagnostic Image

When Qwen rejects a reference image, create a diagnostic copy with these characteristics:

PNG or JPEG.
Standard RGB color.
No animation.
No password protection.
No unusual metadata dependency.
Approximately 1024 pixels on the longest side.
Clearly below the visible upload-size limit.

If the clean copy works, the failure was probably related to the original file’s encoding, size, dimensions, animation, transparency, metadata, or corruption rather than the prompt itself.

Documented Qwen-Image API Output Limits

The following table summarizes current Alibaba Cloud Model Studio specifications. It does not prove that Qwen Create exposes the same Model IDs or controls.

Model family	Output format	Output count	Resolution behavior	Important note
`qwen-image-2.0-pro` and `qwen-image-2.0`	PNG	1–6	Total pixel count must fall between the equivalent of 512×512 and 2048×2048 pixels	Custom width and height are supported within the total-pixel range.
`qwen-image-max`	PNG	1	Uses documented preset sizes	Setting the output count above one returns an error.
`qwen-image-plus`	PNG	1	Uses documented preset sizes	Supports the documented asynchronous Qwen-Image workflow.
`qwen-image-3.0-pro`	PNG	1–6	Total pixel count must fall between the equivalent of 512×512 and 2048×2048 pixels	The current documentation describes this API model as limited preview.

The current Qwen-Image API reference also states that the Qwen-Image 2.0 series accepts prompts of up to approximately 1,300 tokens. Other listed Qwen-Image models accept up to approximately 800 tokens, and excess content can be truncated.

Prompt truncation does not always return an obvious failure. The model may generate an image while ignoring instructions placed near the end. Put the subject, composition, required text, and most important exclusions near the beginning.

Qwen Image Export Formats

The documented Qwen-Image API output is PNG. That makes PNG the safest master format for API-generated images because it preserves the exact returned file without an additional lossy conversion.

The public Qwen Create page does not publish a universal promise that every consumer-interface download will always use the same extension. After downloading an image, verify the actual file type rather than trusting only the filename shown by the browser.

Should You Convert a Qwen Image to WebP?

For websites and WordPress articles, keep the downloaded PNG as the master and create a WebP derivative for publishing.

Download and preserve the original result.
Confirm its dimensions and actual format.
Rename the master using a descriptive filename.
Create a separate WebP copy at the final website dimensions.
Review text, logos, hands, faces, and small details after conversion.
Upload the WebP copy to WordPress while retaining the original locally.

Do not repeatedly convert a JPEG to PNG, then to WebP, and back to JPEG. Multiple transformations can increase file size or introduce additional artifacts without restoring lost detail.

Recommended Image Export by Use Case

Use case	Recommended working format	Recommended published format
WordPress featured image	Original PNG master	WebP at the required dimensions
Infographic with small text	PNG	Lossless or high-quality WebP after visual inspection
Photographic social image	PNG master	High-quality JPEG or WebP
Further image editing	Original PNG	Export a separate final copy
Archive and reproducibility	Original file plus prompt and metadata	Not applicable

Qwen Video Input Formats and Limits

Qwen Create’s public landing page does not provide a complete table of video-upload requirements. For a failing reference video, use a short, standard MP4 or MOV file and remove optional audio, extra references, and unusual encoding before testing again.

The current Wan 3.0 Model Studio API documentation provides the following developer-side examples:

Input video formats: MP4 and MOV.
Input duration: One to 15 seconds per clip, with a maximum combined input duration of 15 seconds.
Input file size: Up to 100 MB per video clip.
Input dimensions: Between 240 and 4096 pixels per side.
Maximum input aspect ratio: 8:1.
Input image formats for video generation: JPEG, JPG, PNG, BMP, and WebP.
Input image file size: Up to 20 MB under the current Wan 3.0 API reference.
Reference audio formats: WAV and MP3.

These are Wan 3.0 API requirements. They should not be represented as guaranteed Qwen Create upload limits unless the interface confirms them.

Documented Wan Video Output Limits

Model or workflow	Resolution	Duration	Output	Notes
`wan2.7-t2v`	720P or 1080P	2–15 seconds	MP4, H.264, 30 fps	Text-to-video with audio support.
`wan2.7-i2v`	720P or 1080P	2–15 seconds	MP4, H.264, 30 fps	Supports first-frame, first-and-last-frame, and continuation workflows.
`wan2.7-r2v`	720P or 1080P	2–10 seconds	MP4, H.264, 30 fps	Reference-to-video with multiple reference types.
`wan2.7-videoedit`	720P or 1080P	2–10 seconds	MP4, H.264, 30 fps	Instruction-based video editing.
`wan3.0-video`	480P, 720P, or 1080P	2–30 seconds without a video input	MP4 result URL; response reports frame rate and dimensions	Supports adaptive duration, common aspect ratios, audio control, and mixed reference media.

Current Wan 3.0 documentation supports these output aspect-ratio controls:

adaptive
16:9
4:3
1:1
3:4
9:16

For tasks containing an input video, the current Wan 3.0 API requires the combined input-video and output-video duration to remain within 30 seconds.

Qwen Video Export Format

Current Model Studio Wan video results are returned through an MP4 URL. Several current Wan models explicitly specify MP4 with H.264 encoding and a 30 fps frame rate.

For a downloaded Qwen video, preserve the original MP4 as the master before editing, compressing, changing its dimensions, or uploading it to a social platform.

Check the Video Before Publishing

Confirm the file opens from beginning to end.
Confirm the duration matches the requested duration.
Confirm portrait and landscape orientation.
Check whether an audio stream exists.
Check dialogue and lip synchronization when audio was requested.
Review the first and final frames for abrupt cuts.
Review text, logos, faces, hands, and product details frame by frame.
Check whether the output includes a watermark.

Do not rely only on the browser preview. A preview can use a lower-resolution stream or fail to reveal a damaged section near the end of the file.

Temporary Result URLs and the 24-Hour Rule

Current Alibaba Cloud documentation states that Qwen-Image result URLs, Wan video URLs, and asynchronous task records are valid for 24 hours. After that period, the asset URL can stop working and a task query can return UNKNOWN.

A temporary URL is not permanent media storage.

Save the task_id and request_id.
Poll the existing task rather than submitting duplicate tasks.
Download the asset immediately after the task succeeds.
Verify the downloaded file.
Move it to permanent storage controlled by you.
Store the permanent URL or file path rather than the temporary provider URL.

If an API image or video existed yesterday but its URL now returns an error, URL expiration may be the intended behavior rather than a new generation outage.

How to Manage Qwen Create Assets

Qwen Create includes an Assets area, but the public landing page does not publish a permanent-retention guarantee for every asset. Treat the Assets area as a working library, not as the only copy of important production media.

For every approved asset, retain:

The original downloaded image or video.
The final prompt.
The source images, video, audio, or document.
The generation date.
The visible model or mode name, when shown.
The selected aspect ratio, resolution, and duration.
The edited and compressed derivatives.
Any usage-rights or client-approval record required for the project.

Common Qwen Create Errors

Error or symptom	Likely meaning	First action
Limit reached or quota exceeded	The current account allowance, plan quota, or API rate limit has been reached	Read the exact notice, stop duplicate submissions, and wait for the stated reset or use an approved API plan.
Task stuck at a high percentage	The job may still be queued, processing, timing out, or failing to update in the browser	Save the prompt, check Assets, and avoid creating repeated copies of the same task.
Generation failed	Invalid input, moderation, capacity, account access, or internal model failure	Test one simple compliant task with no reference media.
Reference file rejected	Unsupported format, excessive size, invalid dimensions, corrupted file, or unreadable URL	Re-export as PNG, JPEG, WebP, MP4, or MOV and test a smaller local copy.
Video duration is invalid	The requested duration is outside the selected model or mode’s range	Use a short standard duration such as five seconds for diagnosis.
Invalid aspect ratio or image size	The width, height, ratio, or total pixel count is unsupported	Use a standard square, landscape, or portrait setting.
DataInspectionFailed	The input or generated output triggered a content-safety check	Remove or rewrite non-compliant content and retry.
IPInfringementSuspect	The prompt or reference media is suspected of involving intellectual-property infringement	Remove protected characters, logos, artwork, or other potentially infringing elements.
429 or throttling	Requests are arriving too quickly or the allocated quota is exhausted	Reduce concurrency and retry with backoff.
500 or 503	Temporary provider, model, or serving-capacity problem	Retry a limited number of times and preserve the request ID.
Task status UNKNOWN	The task ID is wrong, belongs to another region, or has expired	Check the original task response and submission time.
Result disappeared	The preview, asset, or temporary result URL may no longer be available	Check Assets and your local downloads; API users may need to regenerate if the 24-hour URL expired.
Download button does nothing	Browser state, extension, blocked storage domain, expired link, or incomplete result	Open the asset directly and test another browser or network.

60-Second Qwen Create Diagnostic Workflow

Copy and save the prompt before refreshing the page.
Confirm that you are in AI Image or AI Video, not normal chat.
Start a new task rather than continuing inside a broken task.
Remove all reference images, videos, audio, and advanced settings.
Use a short English diagnostic prompt.
Generate one standard-size image or one short 720P video when those options are available.
Check the visible quota or account message.
Check Assets before submitting a duplicate task.
Try a private browser window or another supported browser.
For the API, record the Model ID, region, endpoint, HTTP status, error code, request ID, and task ID.

If the simple task succeeds, add the original resolution, duration, references, audio, text, and advanced controls back one at a time. The first element that causes the failure identifies the relevant limit.

Qwen Create Limit Reached

A Limit Reached message does not tell you which technical limit was reached unless the interface provides additional detail.

It can refer to:

A daily or rolling account allowance.
A plan credit balance.
A limit specific to AI Video.
A concurrent-task limit.
An API requests-per-minute limit.
An API capacity or allocation quota.

Do not repeatedly press Generate. Duplicate submissions can make the queue harder to understand and can create multiple billable API tasks.

Record the complete message, check whether Qwen states a reset time, and compare the image and video sections separately. For a business workflow requiring predictable volume, review the official Model Studio rate-limit and pricing documentation rather than depending on a consumer quota.

Qwen Image or Video Task Stuck

A task that remains at a high percentage is not necessarily still generating. The progress indicator may reflect queue state, inference progress, post-processing, safety checks, upload processing, or a browser update problem.

In Qwen Create

Save the prompt and source files.
Wait once rather than refreshing repeatedly.
Open Assets in another tab and look for a completed result.
Check whether another simple task completes.
Test another browser or trusted network.
Do not submit several identical video tasks simultaneously.

In Model Studio API

Current Wan video APIs are asynchronous. The normal state sequence is:

PENDING → RUNNING → SUCCEEDED or FAILED

Poll the same task_id at a reasonable interval. Current Wan 3.0 documentation suggests an interval such as 15 seconds and notes that video generation commonly takes several minutes.

Creating another task every time the first task remains PENDING wastes resources and can trigger additional throttling.

Qwen Video Generation Error

When Qwen video generation fails, reduce the task to the smallest valid version.

Use text-to-video without an input image, video, or audio.
Use one subject and one scene.
Select a standard 16:9 ratio.
Select 720P when available.
Select a short duration such as five seconds.
Disable optional audio and advanced controls.
Use a harmless prompt without brands, real people, or protected characters.
Generate only one video.

If the baseline succeeds, add the first frame, last frame, reference video, reference audio, longer duration, and higher resolution separately.

Conflicting Video Modes

A video API can reject a request when mutually exclusive inputs are combined. For example, the current Wan 3.0 documentation shows an InvalidParameter error when a reference mode and first-frame or last-frame mode are submitted together.

Select one generation path per diagnostic request:

Text-to-video.
First-frame image-to-video.
First-and-last-frame video.
Reference-to-video.
Video continuation.
Video editing.

Reference Image or Video Rejected

A file can have a familiar extension and still be rejected because its internal encoding is damaged or different from the extension.

Symptom	Likely cause	Fix
Invalid file format	Unsupported container or mismatched extension	Re-export the asset using a standard encoder.
Failed to decode image	Corrupted image or unsupported internal encoding	Open and export a clean PNG or JPEG copy.
Invalid image size	Dimensions, ratio, or total pixel count outside the selected route’s limits	Resize to a standard ratio and moderate dimensions.
Video file too large	File exceeds the selected model or upload limit	Trim, resize, or recompress the video.
Video duration invalid	Clip is longer or shorter than the selected mode allows	Create a short five-second diagnostic clip.
Download media failed	The provider cannot access the supplied URL	Use a public, unexpired HTTPS URL or a supported upload method.
Expired signed URL	The input URL was temporary and expired before processing	Generate a new URL with sufficient validity.
Transparent image behaves incorrectly	The selected video workflow may not support an alpha channel	Flatten the image onto a solid background and export again.

For upload-specific diagnosis, use the separate Qwen file upload troubleshooting guide.

Content Moderation and IP Errors

Image and video generation requests can be checked before and after generation.

DataInspectionFailed

DataInspectionFailed means that the input or generated output was blocked by a content-safety check.

Review the prompt and every reference asset.
Remove sexual, violent, illegal, deceptive, or otherwise restricted elements.
Replace ambiguous wording with neutral, literal instructions.
Test a harmless baseline prompt.
Do not attempt to evade the safety system.

IPInfringementSuspect

IPInfringementSuspect indicates that input data may involve intellectual-property infringement.

Remove unnecessary references to protected characters, brand logos, identifiable artwork, copyrighted media, or instructions to replicate a living artist’s work exactly. Use original descriptions and assets for which you have appropriate rights.

Qwen Download or Export Not Working

An export problem can occur even after generation succeeds.

Open the completed asset rather than downloading from a still-loading preview.
Check the Assets area for another copy.
Allow downloads for the Qwen Create domain in the browser.
Temporarily disable extensions that block scripts, storage domains, or downloads.
Try another supported browser.
Try another trusted network.
For an API URL, check whether its 24-hour validity period has ended.
Confirm that the downloaded file contains media rather than an HTML error page.

A file named image.png can still contain an error document if the download request was redirected or denied. Check the content type and open the file before adding it to a production workflow.

Common Model Studio Error Codes

Code or status	Meaning	Recommended action
`InvalidParameter`	A parameter is missing, conflicts with another mode, or falls outside the allowed range	Compare the request with the exact model’s current API reference.
`InvalidApiKey`	The key is missing or invalid	Use the correct regional key and keep it server-side.
`AccessDenied`	The account lacks permission, preview access, model activation, or workspace access	Check the Model Gallery, workspace, region, and RAM permissions.
`ModelNotFound`	The Model ID is wrong, unavailable in the region, or not activated	Copy the exact Model ID from the regional model list.
`InvalidFile.Format`	The file type is unsupported	Convert it to a documented image, video, or audio format.
`InvalidImage.ImageSize`	The image dimensions or ratio are outside the accepted range	Resize to a standard supported resolution.
`DataInspectionFailed`	Content-safety inspection blocked the input or output	Modify the content and retry.
`IPInfringementSuspect`	The input may involve intellectual-property infringement	Remove infringing or protected content.
`429 Throttling`	Request rate, burst rate, or allocated quota was exceeded	Queue requests and retry using exponential backoff and jitter.
`500 InternalError`	Temporary internal model or service failure	Retry a limited number of times and retain the request ID.
`503`	The model or serving capacity is temporarily unavailable	Retry later or use a tested approved fallback.
`UNKNOWN`	The asynchronous task cannot be found or has expired	Check the task ID, region, workspace, and submission time.

For a broader developer error reference, see Qwen API Errors and Troubleshooting.

How to Verify the Real Export Format

Use the file content rather than the filename alone.

Linux or macOS

file qwen-output.png
file qwen-output.mp4

ImageMagick

magick identify -verbose qwen-output.png

FFprobe

ffprobe -v error \
  -show_entries format=format_name,duration,size \
  -show_entries stream=codec_type,codec_name,width,height,r_frame_rate \
  -of json \
  qwen-output.mp4

The FFprobe result can confirm:

The container format.
The video codec.
The width and height.
The duration.
The frame rate.
Whether an audio stream exists.

Python Example: Save a Temporary Qwen API Result

This example downloads a temporary PNG or MP4 URL, checks the content type and size, and writes it to a local file. It is intended for API result URLs, not for bypassing a website’s normal access controls.

from __future__ import annotations

import os
from pathlib import Path

import requests


CONTENT_TYPE_EXTENSIONS = {
    "image/png": ".png",
    "image/jpeg": ".jpg",
    "image/webp": ".webp",
    "video/mp4": ".mp4",
    "video/quicktime": ".mov",
}


def download_qwen_asset(
    url: str,
    output_stem: Path,
    max_bytes: int = 600_000_000,
) -> Path:
    """Download a temporary Qwen API result with basic validation."""

    with requests.get(
        url,
        stream=True,
        timeout=(10, 120),
        allow_redirects=True,
    ) as response:
        response.raise_for_status()

        content_type = response.headers.get(
            "Content-Type",
            "",
        ).split(";")[0].strip().lower()

        extension = CONTENT_TYPE_EXTENSIONS.get(content_type)
        if extension is None:
            raise ValueError(
                f"Unexpected content type: {content_type or 'missing'}"
            )

        declared_size = response.headers.get("Content-Length")
        if declared_size and int(declared_size) > max_bytes:
            raise ValueError("The asset exceeds the configured size limit.")

        output_path = output_stem.with_suffix(extension)
        temporary_path = output_path.with_suffix(
            output_path.suffix + ".part"
        )

        total = 0
        with temporary_path.open("wb") as file_handle:
            for chunk in response.iter_content(chunk_size=1024 * 1024):
                if not chunk:
                    continue

                total += len(chunk)
                if total > max_bytes:
                    temporary_path.unlink(missing_ok=True)
                    raise ValueError(
                        "The downloaded asset exceeded the size limit."
                    )

                file_handle.write(chunk)

        temporary_path.replace(output_path)
        return output_path


if __name__ == "__main__":
    result_url = os.environ["QWEN_RESULT_URL"]
    saved_path = download_qwen_asset(
        url=result_url,
        output_stem=Path("qwen-generated-asset"),
    )
    print(f"Saved: {saved_path.resolve()}")

Install the required dependency with:

pip install requests

After downloading, inspect the image or video using the verification tools above before moving it into permanent storage.

Recommended Qwen Create Export Workflow

Save the prompt. Keep the final prompt outside the browser session.
Save source assets. Keep original images, video, audio, and documents.
Download immediately. Do not rely on a temporary preview or result URL.
Verify the file. Confirm format, dimensions, duration, codec, and audio.
Preserve a master. Keep the original PNG or MP4 unchanged.
Create derivatives. Produce WebP, JPEG, resized, compressed, or edited copies separately.
Use descriptive filenames. Avoid random browser-generated names in the final library.
Record metadata. Save the date, visible mode, dimensions, duration, prompt, and source rights.
Review manually. Inspect every important detail before publishing.
Back up approved assets. Store them in a controlled local or cloud repository.

When to Use Qwen Create, API, or Local Models

Requirement	Recommended route	Reason
Occasional manual image or video creation	Qwen Create	Fast interactive workflow with AI Image, AI Video, and Assets.
Prompt experimentation inside normal chat	Qwen Studio	Convenient for combining chat, files, research, and generation.
Repeatable dimensions and automated downloading	Model Studio API	Model-specific parameters, response metadata, task IDs, and documented formats.
High-volume application workflow	Model Studio API	Better control over queueing, retries, logging, storage, and usage.
Maximum infrastructure and data-flow control	Local or self-hosted models	The operator controls the runtime, storage, network, and export pipeline.
No-code creative experimentation	Qwen Create	No custom integration is required.

The easiest interface is not always the most predictable production system. Use Qwen Create for interactive creation and a documented API when exact limits, repeatable parameters, automatic storage, and error handling are required.

Frequently Asked Questions

Does Qwen Create have a daily limit?

Qwen does not currently publish one public Qwen Create quota table that guarantees the same daily allowance for every account. Check the counter, account notice, plan information, or Limit Reached message shown in your current interface.

What image formats does Qwen Create support?

The public Qwen Create landing page does not list a complete consumer-interface format table. For troubleshooting, use a clean PNG, JPEG, or WebP image. Current Qwen Image 3.0 API documentation supports JPG, JPEG, PNG, BMP, TIFF, WebP, and GIF reference images, subject to model-specific size and dimension limits.

What format does Qwen export images in?

Current official Qwen-Image APIs return PNG results. Qwen Create’s consumer-interface export behavior should be verified from the actual downloaded file because its public page does not publish a universal export-format guarantee.

Can Qwen export WebP images?

The documented Qwen-Image API output is PNG rather than WebP. Download the original PNG, preserve it as the master, and create a separate WebP derivative for your website.

What format does Qwen export videos in?

Current Model Studio Wan video workflows return MP4 results. Several current Wan families specify H.264 encoding and 30 fps. Verify the Qwen Create browser download separately because the public product page does not guarantee one export specification for every mode.

How long can a Qwen video be?

There is no single duration that applies to every Qwen Create and Model Studio route. Current Wan 2.7 text-to-video and image-to-video APIs support up to 15 seconds, while Wan 3.0 supports up to 30 seconds without an input video. The duration available in Qwen Create must be checked in the current interface.

Can Qwen Create export 4K images or video?

Do not assume that Qwen Create provides universal 4K export. Current Qwen-Image 2.0 and Qwen Image 3.0 API documentation limits output by total pixel count up to the equivalent of 2048×2048 pixels. Current Wan video APIs discussed in this guide provide up to 1080P. Other Alibaba image models can have different limits, but those limits do not automatically apply to Qwen Create.

Why is Qwen video generation stuck at 99%?

The job may be queued, processing, completing post-processing, waiting for a safety check, or failing to update in the browser. Save the prompt, check Assets, avoid duplicate submissions, and test a short simple task. API users should continue polling the original task ID.

Why does Qwen say video generation failed?

Common causes include an unsupported input, invalid duration or ratio, conflicting generation modes, content moderation, an expired reference URL, an account limit, rate limiting, or a temporary server error. Test a five-second 720P text-only video with no reference media when those options are available.

Why did my Qwen image or video disappear?

Check the Qwen Create Assets area and your browser downloads. For Model Studio APIs, image URLs, video URLs, and task IDs are currently valid for 24 hours. After expiry, the asset must come from your own saved copy or be generated again.

Does Qwen Create add a watermark?

Watermark behavior can depend on the product mode and current interface. Model Studio image and video APIs expose model-specific watermark parameters, but those API defaults should not be treated as proof of Qwen Create UI behavior. Inspect the final downloaded asset.

Can Qwen Create generate several images at once?

The answer depends on the interface and model. Current Qwen-Image 2.0 and 3.0 API models can support one to six outputs, while Qwen Image Max and Plus API routes are documented with one output. Check the current Qwen Create control rather than assuming the API setting is exposed.

Is Qwen Create the same as Alibaba Cloud Model Studio?

No. Qwen Create is an interactive creative product. Alibaba Cloud Model Studio is a developer platform exposing specific Model IDs, regional endpoints, API keys, pricing, quotas, and response objects. Their limits and workflows should be documented separately.

Where can I find deeper Qwen image troubleshooting?

Use the dedicated guide: Qwen Image Generation Not Working. It covers Qwen Studio, Qwen Create, Model Studio API errors, third-party providers, and local Qwen-Image failures.

Conclusion

The most important Qwen Create limit is not one universal number. Limits are attached to the specific product route, account, task type, input asset, selected mode, model, and output settings.

Use the current Qwen Create interface as the source for your account’s visible quota.
Do not transfer Model Studio API limits to Qwen Create without confirmation.
Current Qwen-Image API outputs are PNG.
Current documented Wan video outputs are MP4.
API image and video result URLs expire after 24 hours.
Preserve the original file before converting or compressing it.
When a task fails, simplify the task before changing several settings at once.
Keep task IDs, request IDs, prompts, source assets, and downloaded masters for important work.

For occasional creative work, Qwen Create offers the simplest route. For production systems that require fixed parameters, automation, predictable storage, and structured error handling, use a documented Model Studio API workflow.

Main Sources Used

Last verified: August 24, 2026
Evidence status: Documentation-Verified, Product-Surface Verified, and Editorial Analysis