Home / Reviews / AI Video Generator / Qwen AI
Qwen logo
AI Assistant Multimodal Chat

Qwen AI Review

qwen.ai · by Alibaba Cloud

Qwen is Alibaba Cloud's AI assistant, a free multimodal chatbot that pairs text, image, video, and agent tools in one web app at qwen.ai. The free tier runs capable chat and image models, with guest access for core tasks and no subscription required. It suits people who want a broad, low-cost alternative to the $20 assistants and do not mind a provider based in China.

Reviewer Score
8.0
/10
Top Pick
Strong free multimodal core
held back by one feature
Translation Quality8.5
Reasoning on Fast8.5
Deep Research8.0
Localization Fit6.0
Podcast Playback4.5
Type
Web AI assistant
Access
Guest or free account
Free Chat Model
Qwen3.7-Plus
Free Image Model
Qwen-Image 3.0
Age Requirement
18 and over
Paid From
$2.12 /mo, Lite

What Qwen Does

01 . Overview

Qwen is the consumer face of Alibaba Cloud's Qwen model family, served as a web assistant at qwen.ai (formerly chat.qwen.ai). It runs text chat, image and video generation, document work, web search, and agent-style tasks from a single interface, and most of it is free.

The positioning is reach for the price. Where ChatGPT, Gemini, and Claude sell their deeper features at around $20 a month, Qwen folds image generation, Deep Research, and agent tools into a free tier that even guests can touch. Paid plans exist for heavier use, opening at a few dollars a month.

What Happened When We Tested It

02 . 6 tests covering translation, reasoning, research, and media generation

We ran Qwen on qwen.ai with a free account, a few tasks as a guest and the rest after logging in. Each test below carries the exact prompt typed and the output Qwen returned, from a casual translation through a text-on-image poster.

An age gate stops a guest translation

Test 01 . Access and gating
★ Prompt Used

"Translate this into Chinese (Simplified), Arabic, Spanish and French, keep it natural: Our new product launches next week. Stay tuned for exclusive offers."

Qwen translation prompt typed as a guest on the Qwen3.7-Plus model, asking for a sentence translated into four languages
The translation prompt, typed as a guest on Qwen3.7-Plus.
Qwen age-check screen returning Age Requirement Not Met after an under-18 age was entered, restricting the chatbot to users 18 and over
An age under 18 returns "Age Requirement Not Met", limiting Qwen to adults.
First impression: opening qwen.ai without signing in and sending a translation did not go straight to an answer. Qwen paused to ask for an age, and an entry below 18 returned an "Age Requirement Not Met" notice. The gate applies even to guests, before a task as ordinary as translating a marketing line.

Four scripts, all reading naturally

Test 02 . Translation output
Qwen output showing the marketing sentence translated into Simplified Chinese, Arabic, Spanish and French, each line phrased naturally
The four-language output, produced with no login once past the age step.
What stood out: past the gate, the translation returned in Simplified Chinese, Arabic, Spanish, and French. The lines read naturally rather than word for word, catching the tone of a product teaser instead of a literal rendering. Clean output across four scripts on a guest session made for a strong opening.

Reasoning on Fast, and the free wall

Test 03 . Reasoning and gating
★ Prompt Used

"A bat and a ball cost $1100 together. The bat costs $1000 more than the ball. How much is the ball? Show your reasoning step by step."

Qwen bat-and-ball prompt on Fast mode, with Upload attachment and Create Video greyed out and marked Log in
The bat-and-ball prompt on Fast mode, with Upload attachment and Create Video greyed behind "Log in".
Qwen answer giving $50 for the ball, showing each reasoning step, verifying the total back to $1100, and noting why people guess $100
Qwen's answer: $50, with a verification step and a note on why people guess wrong.
The key finding: the "+" menu told its own story first. Upload attachment and Create Video sat greyed out, each marked "Log in", so guests get text chat while the heavier tools wait behind an account. Left on Fast, Qwen still got the trick right: it returned $50, showed each step, verified the total back to $1100, and explained why most people guess $100, all without the slower Thinking mode.

Deep Research asks before it runs

Test 04 . Deep Research
★ Prompt Used

"Best budget smartphones under $600 in USA in 2026. Give a structured report with pros, cons and sources."

Qwen Deep Research prompt for a US budget-phone report, with sample suggestion prompts below that lean toward Chinese topics
The Deep Research prompt. The sample suggestions below lean toward Chinese topics.
Qwen asking three clarifying questions before Deep Research begins, with a research plan laid out in the right pane
Qwen asked three clarifying questions first, then laid out a research plan.
Qwen Deep Research finished in 7 minutes 21 seconds with a 16-page report and a Create button offering a webpage or a podcast
Deep Research finished in 7 minutes 21 seconds. The Create button offers a webpage or a podcast.
How it performed: one detail landed before the run even started. The suggested example prompts skewed Chinese, covering the film Ne Zha 2 and a week in Hong Kong planned from Beijing. Qwen then asked three clarifying questions, which most free tools skip, and a planning pane set the scope. The run took 7 minutes 21 seconds and produced a 16-page report titled "Beyond Megapixels", with a Create button offering a live webpage or a podcast.

The podcast that starts 30 minutes in

Test 05 . Audio output
Qwen queuing a podcast generation job from the Deep Research report, with a message that it will notify on completion
Creating the podcast. Qwen queued the job with a completion notice.
The finished Qwen podcast listing a length of 35 minutes 2 seconds but starting playback at the 30-minute mark on every replay
The finished podcast lists a 35:02 length but starts playback at 30:00 on every replay.
Where it lands: the podcast built without error and queued with a completion notice. The finished file lists a length of 35 minutes 2 seconds, yet every press of play opened it at the 30-minute mark. Replaying many times gave the same jump each run, leaving the first 30 minutes unreachable from the player. The feature generates; the playback does not work.

Text on an image, spelled right

Test 06 . Image generation
★ Prompt Used

"A motivational poster: mountain background at sunrise, with the exact bold white text KEEP CLIMBING across the top."

Qwen Create Image screen on the free account running Qwen-Image 3.0, with the sharper Pro model carrying a VIP tag
The image prompt on Create Image. The free account uses Qwen-Image 3.0; the Pro model carries a VIP tag.
Qwen image result showing KEEP CLIMBING spelled correctly in bold white text across a sunrise mountain range, with a Qwen watermark in the lower corner
The result. KEEP CLIMBING is spelled correctly and sharp, with a Qwen watermark in the lower corner.
What this confirms: the text-on-image poster is the job that trips up most image tools, and the free account, running Qwen-Image 3.0, placed KEEP CLIMBING across a sunrise mountain range with correct spelling and clean edges. This is where Qwen earns its reputation. Two limits apply to the free result: a "Qwen" watermark sits in the lower corner, and the sharper Pro model is paywalled behind a VIP tag.
★
Guest translation, correct reasoning, and in-image text that spells right

Testing confirmed a strong core: a four-language translation that read naturally on a guest session, a trick reasoning question answered correctly on Fast mode with a verification step, a Deep Research run that asked clarifying questions before producing a 16-page report, and a poster whose KEEP CLIMBING text came out correctly spelled.

!
The podcast plays from the 30-minute mark

The Deep Research podcast built cleanly but would not play from the start. A file listed at 35 minutes opened at 30:00 on every attempt, leaving the first half hour unreachable. A feature that generates but cannot be heard in full is not yet usable.

How this review was put together. This review is based on hands-on use of Qwen at qwen.ai in November 2026, a few tasks as a guest and the rest on a free account, with no payment. Testing covered translation, reasoning on Fast mode, Deep Research, the podcast output, and image generation. Pricing was checked against Qwen's own in-app plan screen rather than a third-party listing.

10-Point Feature Review

03 . Ten scores, weighted toward core chat and reasoning
Feature Scores at a Glance
Translation quality
8.5
Reasoning on Fast
8.5
Deep Research
8.0
In-image text
8.0
Guest access
7.5
Multimodal breadth
7.5
Free-tier clarity
6.5
Localization fit
6.0
Age-gate friction
6.0
Podcast output
4.5

Pros and Cons

04 . What a real user notices
+What works
  • Strong multilingual output: translations read naturally across different scripts.
  • Reliable reasoning on Fast: it handled a classic trap question with a verification step.
  • Deep Research scopes first: it asks clarifying questions before the run begins.
  • Broad toolset in one place: chat, image, video, agent, and research together.
−What grates
  • Podcast playback is broken: generated audio opens partway in, not at the start.
  • Age gate blocks guests: under-18 users are stopped before even a translation.
  • Heavier tools need login: upload and video sit behind an account for guests.
  • Region-skewed defaults: suggested prompts lean to Chinese topics for a US user.

Pricing Breakdown

05 . A free tier, then three paid steps
Feature Free$0 Lite$2.12 /mo Plus$20.23 /moRecommended Pro$112.93 /mo
Model access Standard limits Higher limits Professional limits
Agent & Qwen Code Basic agent Basic (same as Free) 5× Plus, shared
Extra video credits Daily bonus only +400 / month +10,500 / month
Image generation Everyday use Higher limits High-volume
Best for Trying it Light everyday Intensive use

The tier math: two things are worth knowing before you upgrade. The USD amounts are oddly precise, which points to conversion from a yuan base rate, so the exact figure can drift with the exchange rate. And the free tier already includes chat, image, video, and agent features at standard limits, so the paid steps mostly buy higher usage and extra monthly credits rather than new abilities, until Plus adds Qwen Code and a shared agent pool. The jump from Plus to Pro is steep, roughly five times the usage for more than five times the money.

Qwen consumer pricing grid showing Free, Lite, Plus and Pro plans with monthly prices and feature allowances
Qwen's four consumer plans, from the free tier up to Pro.

Pricing verified November 16, 2026 from Qwen's in-app plan screen. The USD figures appear converted from a yuan rate, so confirm the current amount on the Qwen plan page before subscribing.

Qwen vs the Top 4 Alternatives

06 . How it stacks up on price and origin
QQwen GChatGPT GGemini CClaude DDeepSeek
MakerAlibaba (China)OpenAI (US)Google (US)Anthropic (US)DeepSeek (China)
Free tier✓✓✓✓✓
Entry paid price$2.12/mo (Lite)$20/mo (Plus)$19.99/mo (AI Pro)$20/mo (Pro)Free only
Best ForCheap step-up, multimodalMost polished all-rounderGoogle Workspace usersWriting and coding depthStrong fully-free option

Where it fits: Qwen's case against the $20 crowd is price and breadth, since the paid step-up starts far lower and image plus video generation sit inside the free chat. ChatGPT stays the most polished all-rounder, Gemini wins if you live in Google Workspace, and Claude leads for writing and coding depth. DeepSeek is the other strong free option, and like Qwen it is China-based, a point some buyers weigh.

· The Verdict ·
8.0/10
A standout free multimodal pick, if you skip the podcast

Use Qwen if you want a free, multimodal assistant that handles translation and reasoning well and gets in-image text right, and you like having the cheapest paid step-up in its class to grow into.

Skip Qwen if you need reliable audio output or a US-based provider. ChatGPT and Claude are the safer all-rounders at $20, and Gemini fits Google Workspace users.

Category RankTop Pick
Compared To4 Alternatives
Entry PriceFree / $2.12 mo
Free TierYes, guest OK

Discussion

Join the discussion and share your perspective.