Use the official Qwen Download page to get Qwen for Android, iPhone, iPad, Windows, or macOS. You can also use Qwen in a web browser without installing an application. If you want to run a model locally or install the Qwen Code development agent, use the separate official sources explained below rather than downloading a consumer app.
Safety rule: begin at qwen.ai/download, even when you plan to install from an app store. Do not download Qwen from an APK mirror, software-download portal, browser extension, file-sharing service, search advertisement, or unofficial GitHub repository.
Which Qwen download do you need?
| Your goal | Correct option | Official source |
|---|---|---|
| Use the normal Qwen assistant without installing anything | Qwen web | Qwen Chat |
| Install Qwen on an Android phone or tablet | Official Android app | Qwen Download or the verified Google Play listing |
| Install Qwen on an iPhone or iPad | Official iOS or iPadOS app | Qwen Download or the verified App Store listing |
| Install the consumer desktop app on a PC | Qwen for Windows | Official Qwen Download page |
| Install the consumer desktop app on a Mac | Qwen for macOS | Official Qwen Download page |
| Use Qwen on Linux without local setup | Qwen web | Qwen Chat |
| Run a Qwen model on your own hardware | Official model weights or a compatible quantization | Qwen on Hugging Face or Qwen on ModelScope |
| Use a Qwen coding agent in a terminal or editor | Qwen Code | Official Qwen Code repository |
| Build an application through an API | Qwen API access | Qwen API Platform; no consumer-app download is required |
Installing the Qwen desktop or mobile app is not the same as downloading a Qwen model. The consumer app provides access to Qwen’s online assistant, while model weights are large files intended for local runtimes, development environments, or self-hosted inference.
How to verify an official Qwen download
A copied name or logo does not prove that an app is official. Use the following checks before installing anything:
- Start from the official domain. Type https://qwen.ai/download directly or reach it from qwen.ai.
- Follow Qwen’s own platform link. Let the official page send you to Google Play, the App Store, or the current desktop installer.
- Check the mobile app name. The current international listing is named Qwen – Alibaba AI Assistant.
- Check the Android identity. The verified Google Play package is
com.tongyi.intl. The listing displays Qwen Team as the developer and identifies NTH POWER GLOBAL TECH SINGAPORE PTE. LTD. in its developer information. - Check the Apple developer. The verified App Store listing identifies NTH POWER GLOBAL TECH SINGAPORE PTE. LTD. as the developer.
- Inspect a desktop installer before running it. Confirm that the file came directly from Qwen’s official page and that the operating system shows a valid, expected publisher or developer identity.
- Stop if the identity is unexpected. Do not bypass an unknown-publisher warning, disable antivirus protection, turn off SmartScreen, remove macOS quarantine protection, or override Gatekeeper for an installer whose source or signer cannot be verified.
- Avoid static installer links copied from another site. Versions, architectures, and file URLs can change. The official landing page is the safer evergreen route.
Store descriptions and privacy declarations are provided by the developer and can change. They are useful verification signals, but they are not independent security audits.
Official Qwen download options
As of October 13, 2026, Qwen’s official site presents five consumer access routes: Web, iOS, Android, macOS, and Windows. It does not list a native Linux consumer application in the same platform selector.
| Version | Best suited for | Important limitation |
|---|---|---|
| Web | Immediate access from a supported browser without installation | Requires browser access and should not be assumed to match every app feature |
| Android | Mobile chat, voice, photos, camera-based tasks, files, and notifications | Availability and compatibility can vary by device and Google Play region |
| iPhone and iPad | Mobile use through the official App Store application | The store region, operating-system requirement, size, and available purchases can vary |
| Windows | A dedicated desktop experience and work involving screen or file context | Use the system compatibility information displayed by the current official download page |
| macOS | A dedicated Mac desktop experience | Use the dedicated macOS download rather than treating the iPhone or iPad listing as the main Mac version |
Qwen’s product overview positions its mobile experience around on-the-go and voice use, while the desktop experience can work with material such as email, screenshots, files, and information on the screen. Exact capabilities can differ by platform, version, account, plan, and region.
How to download Qwen on Android
- Open Qwen’s official download page on the Android device.
- Select the Android option and follow the link provided by Qwen.
- On Google Play, confirm the name Qwen – Alibaba AI Assistant.
- Confirm that the package URL ends in
id=com.tongyi.intl. - Check that the listing shows Qwen Team and the official developer information.
- Review whether Google Play marks the device as compatible.
- Review the current permissions and data-safety information before installation.
- Select Install, open the app, and follow the current sign-in flow.
- Test the app with a short, non-sensitive prompt before uploading files, enabling the microphone, or granting access to photos and the camera.
The Google Play listing checked on October 13, 2026 described Qwen as Alibaba’s official AI app, displayed in-app purchases, and showed a last-update date of September 30, 2026. The listing also publishes [email protected] for app issues and feature suggestions.
Should you download a Qwen APK?
The verified international store route is the Google Play package com.tongyi.intl. Do not install an APK from a mirror merely because it uses Qwen’s name or icon.
If Qwen’s official download page itself presents a direct Android package for your device or region, use only the first-party link provided during that visit and verify the package identity before installation. Direct package URLs can change, so do not rely on an APK link copied from a forum, tutorial, file host, or software portal.
If Google Play says the app is unavailable, that can mean the device, operating-system version, account region, store region, or rollout is unsupported. Use Qwen on the web rather than bypassing the restriction with an unverified package.
How to download Qwen on iPhone or iPad
- Open the official Qwen Download page.
- Select the iOS option and let Qwen send you to the App Store.
- Confirm the app name Qwen – Alibaba AI Assistant.
- Confirm that the developer is NTH POWER GLOBAL TECH SINGAPORE PTE. LTD.
- Review the compatibility, privacy, and in-app-purchase information for your own App Store region.
- Select Get and complete Apple’s installation process.
- Open Qwen and follow the current sign-in or onboarding flow.
- Begin with a non-sensitive test before granting optional photo, microphone, camera, or notification permissions.
The Singapore App Store listing checked on October 13, 2026 required iOS 16.0 or later for iPhone and iPadOS 16.0 or later for iPad. It listed version 2.0.5, dated September 30, 2026. These details are point-in-time data from one regional store and may change or differ elsewhere.
That mobile listing also showed Apple-silicon Mac compatibility but labeled the app as not verified for macOS. Mac users should therefore use the dedicated macOS route on Qwen’s official download page rather than treating the mobile App Store version as the primary Mac application.
How to download Qwen on Windows
Qwen lists a native Windows option on its official website. Use the official landing page instead of a software-download mirror because the current installer, architecture, and compatibility requirements can change.
- Open qwen.ai/download on the Windows computer.
- Select Windows.
- Read the current compatibility information before saving the file.
- Confirm that the browser is downloading from Qwen’s official domain or a first-party destination reached directly from it.
- After the download finishes, right-click the installer and open Properties.
- Inspect the publisher information and the Digital Signatures tab when present.
- Do not run the installer if the signer is missing, invalid, or unrelated to the official Qwen download flow.
- Run the verified installer and follow its prompts.
- Launch Qwen and complete a short non-sensitive test.
Qwen’s accessible public pages did not expose a stable Windows minimum-version, architecture, installer-size, or filename table during this review. Check the current download page and installer properties instead of relying on a requirement copied from an old article.
If Windows shows an unknown or unexpected publisher, stop. Return to the official page, delete the unverified copy, and download it again. Do not disable SmartScreen or other security protections merely to make the file run.
How to download Qwen on macOS
- Visit the official Qwen Download page on the Mac.
- Select macOS.
- If Qwen presents separate processor builds, choose the one matching the Mac’s processor. Do not guess between Intel and Apple silicon.
- Download the current installer or disk image from the first-party route.
- Open the file and confirm the developer identity shown by macOS.
- Install the app only when the source and developer are expected.
- Open Qwen and grant only the permissions needed for the feature you intend to use.
- Run a non-sensitive test before allowing access to files, screenshots, screen content, microphone, or other data.
Qwen did not expose a stable public table of consumer desktop requirements in the first-party pages available during this review. Confirm the minimum macOS version and processor support on the live download page before installing.
If macOS says the developer cannot be verified, do not immediately override Gatekeeper. Confirm that the download originated from Qwen’s official page and that the developer identity is valid. Redownload the file if the source or signature is unclear.
Can you download Qwen for Linux?
As of October 13, 2026, Qwen’s official consumer download selector listed Web, iOS, Android, macOS, and Windows, but not a native Linux consumer app.
Linux users have three legitimate options:
- Use Qwen in a browser. This is the simplest route for ordinary users.
- Run a Qwen model locally. This requires suitable hardware and a compatible runtime such as Transformers, vLLM, llama.cpp, Ollama, or LM Studio.
- Install Qwen Code. This is appropriate for software-development workflows, not as a replacement for the normal consumer assistant.
An unofficial Linux wrapper may display the Qwen website inside a desktop shell, but that does not make it an official Qwen application. Review its code, publisher, update process, and credential handling separately before using it.
Qwen web versus mobile and desktop apps
| Option | Main advantage | Choose it when |
|---|---|---|
| Web | No installation and the fastest official access route | You use Linux, a shared computer, an unsupported device, or do not need native permissions |
| Mobile app | Convenient access to voice, camera, photos, files, and notifications | You mainly use Qwen from a phone or tablet |
| Desktop app | A dedicated desktop window and closer interaction with files, screenshots, email, or screen context where supported | You use Qwen regularly on Windows or macOS |
| Local model | Control over the selected runtime and model files | You have appropriate hardware and understand local-AI configuration |
Do not assume complete feature parity. A tool available on mobile may be absent on desktop, and a desktop capability may not appear on the web. Availability can also vary by account, subscription, model, region, and rollout.
First launch, account access and synchronization
After installation, use the sign-in method associated with your existing Qwen account. For example, if you originally used a social sign-in option, selecting an unrelated email flow can create confusion about missing history or subscriptions.
- Confirm that the app was reached from Qwen’s official page.
- Use the intended account and sign-in method.
- Test a short chat before uploading files.
- Check whether a test conversation appears on a second device before assuming that all history synchronizes.
- Check file access, generated media, settings, and subscriptions separately; one synchronized item does not prove that everything is shared across platforms.
- Do not enter a password or one-time code into a page opened from an unexpected email, pop-up, extension, or third-party client.
Qwen’s public pages do not provide one complete matrix guaranteeing synchronization of every chat, file, setting, purchase, and feature across web, mobile, and desktop. Verify the specific item that matters to your workflow before deleting a local copy or moving between platforms.
Does the Qwen app work offline?
Installing the consumer Qwen app does not mean that a full Qwen model has been downloaded to the device. Qwen presents the consumer app as an online AI assistant, and its public download pages do not describe the normal mobile or desktop application as a complete offline local-model package.
For genuinely local inference, you need:
- A downloaded Qwen model or compatible quantization.
- A runtime that supports that model format.
- Sufficient storage, RAM, and—in many cases—GPU or unified memory.
- A configuration that does not send prompts to a remote provider.
Even a locally installed model may require internet access for its first download, updates, remote tools, web search, authentication, plugins, or API-backed functions. Local privacy also depends on the runtime, extensions, network settings, logging, and operating environment.
How to download Qwen model weights
Use the official Qwen organization on Hugging Face or the official Qwen organization on ModelScope. These repositories are separate from the consumer Qwen application.
- Define the task. Decide whether you need text generation, vision and text, image generation, audio, speech, embeddings, reranking, or another specialized capability.
- Choose an appropriate model size. A larger or newer model is not automatically suitable for the available hardware.
- Confirm the repository owner. On Hugging Face, an official repository should appear under the Qwen organization.
- Read the model card. Check the intended use, architecture, context support, limitations, runtime instructions, and license.
- Inspect the file list before downloading. Full model weights can require substantial disk space and memory.
- Choose the correct format. Transformers or Safetensors weights, FP8 variants, GGUF files, and MLX conversions serve different runtimes and hardware.
- Verify whether a conversion is official. A community GGUF or MLX conversion can be useful without being an official Qwen release.
- Follow the selected model card. Use its current Transformers, vLLM, SGLang, llama.cpp, Ollama, LM Studio, MLX, or other documented setup.
Use our local Qwen setup guide to match a checkpoint and runtime to your available memory.
Full weights versus quantized files
| Format | Typical use | What to verify |
|---|---|---|
| Transformers or Safetensors | Python, Transformers, vLLM, SGLang, and compatible serving systems | Model card, license, framework version, storage, RAM, and GPU requirements |
| Official FP8 or reduced-precision variant | Compatible accelerators and serving runtimes | Hardware support and whether the repository is under the official Qwen organization |
| GGUF | llama.cpp, Ollama, LM Studio, and related local runtimes | Quantization level, conversion author, source model, context configuration, and runtime compatibility |
| MLX conversion | Local inference on compatible Apple-silicon systems | Conversion source, quantization, model license, and current MLX compatibility |
For example, the official Qwen3.8-27B model card identifies the repository as Qwen-owned, lists an Apache 2.0 license, provides Transformers-format weights, and links to quantizations for compatible local applications. That makes it a useful example of what an official model page looks like, but it is not automatically the right model for every computer.
Do not assume that all Qwen models have the same license or hardware requirements. Verify the selected repository individually.
Qwen Code is a separate download
Qwen Code is Qwen’s separate open-source coding agent for terminal, editor, desktop, browser, and chat workflows. It is not the normal Qwen consumer app.
For the separate coding agent, follow our Qwen Code installation and authentication guide.
The official repository provides standalone installers for Linux, macOS, and Windows, along with npm and Homebrew installation routes. As of October 13, 2026, the documented npm method requires Node.js 22 or later:
npm install -g @qwen-code/qwen-code@latestAfter installation, open a terminal in the intended project and run:
qwenThe official quick start then uses /auth to configure the selected provider. Follow the current repository documentation because authentication options, provider support, and installation methods can change.
Qwen Code’s Linux support does not mean that the normal consumer Qwen desktop app has an official Linux version. They are separate products with different purposes and installation paths.
Permissions and privacy before using the app
Features such as voice, camera input, photo analysis, file upload, screenshots, screen awareness, and notifications may require device permissions. Grant a permission only when you intend to use the related feature.
- Start with a non-sensitive prompt.
- Do not upload passwords, API keys, authentication tokens, identity documents, payment information, confidential customer data, or private business files without an approved reason.
- Allow microphone or camera access only when using voice, image, or video features.
- Review file and photo access before choosing broad library permissions.
- Revoke permissions that are no longer needed.
- Read the current Qwen Privacy Policy and Terms of Service.
- Check organizational policy before installing Qwen on a managed work or school device.
The Google Play listing says its data-safety information is supplied by the developer and may vary by use, region, and age. Apple similarly states that its displayed privacy information is developer-provided and has not been independently verified by Apple. Treat store declarations as disclosures to review, not as absolute privacy guarantees.
A cloud app, local model, API client, and third-party runtime can have different data flows. Downloading open-weight model files does not automatically make the surrounding runtime private, and the consumer app’s privacy terms do not automatically govern an unrelated local or third-party tool.
How to update or uninstall Qwen
Update the mobile app
- On Android, open the verified Google Play listing and select Update when available.
- On iPhone or iPad, use the App Store’s update controls.
- Confirm the developer and package identity before updating an app installed from an unusual source.
Update the desktop app
Use Qwen’s in-app updater when the installed version provides one, or return to the official download page for the current desktop installer. Do not download an “update” offered by a pop-up, email attachment, browser extension, or unrelated website.
Uninstall Qwen
Remove the application through the normal app-management controls on Android, iOS, Windows, or macOS. Uninstalling the app is not the same as:
- Signing out.
- Deleting the Qwen account.
- Deleting server-held chat history.
- Cancelling a paid subscription.
- Removing files downloaded by a local-model runtime.
- Deleting Qwen Code configuration or project files.
Complete each of those actions separately when required, using the current account, store, subscription, or runtime controls.
Fix common Qwen download and installation problems
| Problem | Safest first check | Next action |
|---|---|---|
| The app is unavailable in the store | Check the device, operating-system version, account region, and store region | Use Qwen web; do not switch to an unverified APK or store account |
| Google Play shows no Install button | Look for an incompatibility or regional message | Update the supported operating system if appropriate or use the web version |
| The App Store says the device is incompatible | Compare the device’s iOS or iPadOS version with the current regional listing | Install a supported operating-system update or use Qwen web |
| The Windows installer will not open | Confirm the file came from qwen.ai and inspect its publisher and signature | Delete an incomplete or unverified copy and download it again from the official page |
| Windows shows an unexpected publisher | Compare the source and signer with the official download flow | Stop installation; do not bypass the warning |
| macOS blocks the app | Verify the official source, developer identity, compatibility, and notarization | Redownload from Qwen; do not override Gatekeeper for an unverifiable file |
| The downloaded file appears incomplete | Compare its completed status and available disk space | Delete only the partial file, use a stable connection, and retry from the official page |
| The app opens to a blank screen | Test Qwen web and check for an app update | Restart the app, compare another network, then reinstall only after preserving needed data |
| Sign-in fails after installation | Confirm the official app and the original sign-in method | Test the web login and another network before clearing app data |
| A local model download stops | Check free disk space, network restrictions, and the official repository | Use the official Hub client or runtime’s supported resume process rather than a mirror |
| Only small Git LFS pointer files were downloaded | Check whether the repository was cloned without the required large-file support | Use the model card’s official download method or the Hugging Face hf download client |
| The local runtime cannot load the model | Compare the model format with the runtime and hardware | Select a compatible official file or verified quantization; do not rename an incompatible format |
The qwen command is not found after Qwen Code installation | Restart the terminal, as the official installer instructs | Verify the installation path and follow the current Qwen Code setup guide |
If installation succeeds but sign-in or loading still fails, continue with our Qwen troubleshooting guide.
Before reinstalling an app, clearing its data, or deleting a model cache, preserve any local files, prompts, generated outputs, or configuration you still need.
Frequently asked questions
Where can I download Qwen officially?
Use qwen.ai/download. It is the official landing page for Qwen’s web, iOS, Android, macOS, and Windows routes.
Is Qwen free to download?
The verified Google Play and App Store listings are free to install and display in-app purchases. Free installation does not mean that every plan, model, feature, or usage level is free.
Can I use Qwen without downloading it?
Yes. Open Qwen Chat in a supported web browser. This is also the simplest official route for Linux, ChromeOS, a shared computer, or an unsupported device.
Does Qwen have an official Linux app?
The consumer download selector checked on October 13, 2026 did not list Linux. Linux users can use Qwen web, run an appropriate Qwen model locally, or install the separate Qwen Code agent for development work.
Is there an official Qwen APK?
The verified international Android store package is com.tongyi.intl on Google Play. If Qwen’s own download page presents a direct package for your device or region, use only that first-party route. Do not install an APK from a mirror or file-sharing site.
Does the Qwen desktop app run a model locally?
Qwen’s consumer download pages do not present the normal desktop application as a complete offline local-model bundle. Installing the desktop app is different from downloading model weights from Qwen’s official Hugging Face or ModelScope organization.
Where should I download Qwen model weights?
Use the official Qwen Hugging Face organization or Qwen ModelScope organization. Verify the repository owner, model card, files, format, license, and runtime requirements before downloading.
Are GGUF and MLX Qwen files official?
Some are official, while many are community conversions. Check the repository owner and conversion notes. A community conversion can be legitimate and useful, but it should not be described as an official Qwen release unless it is published or explicitly endorsed by Qwen.
Is Qwen Code the same as the Qwen app?
No. Qwen Code is a development agent for terminal, editor, desktop, browser, and coding workflows. The Qwen consumer app is a general AI assistant. They have separate installation instructions and use cases.
How can I avoid fake Qwen apps?
Start from Qwen’s official download page, confirm the app name and developer, check the Android package ID or Apple app ID, verify desktop signatures, and avoid mirrors, search advertisements, browser extensions, and direct files whose publisher cannot be confirmed.
Sources checked
- Official Qwen Download page for the current consumer platform routes.
- Official Qwen product overview for web, mobile, and desktop positioning.
- Qwen on Google Play for the Android app name, package, developer, update date, support route, and store disclosures.
- Qwen on the Singapore App Store for the iPhone and iPad app, developer, version, and regional compatibility information.
- Official Qwen organization on Hugging Face for first-party model repositories.
- Official Qwen organization on ModelScope for model downloads.
- Official Qwen GitHub organization for Qwen repositories and model projects.
- Official Qwen Code repository for current installation and quick-start instructions.
- Qwen Privacy Policy and Qwen Terms of Service for current service and data terms.
Platform availability, store listings, app versions, system compatibility, installers, model repositories, licenses, and Qwen Code commands were last verified on October 13, 2026. Return to the relevant first-party page before installing because these details can change.