Qwen (chat.qwen.ai) is Alibaba's flagship AI assistant, and its growth trajectory in early 2026 is one of the most striking data points in the global AI market. The app's monthly active users surged from 31 million in January 2026 to 203 million in February — a 554% increase in 30 days — driven by Alibaba's Lunar New Year campaign and the release of Qwen3.5, now ranked among the top open-weight models globally on Hugging Face. It currently sits third globally behind ChatGPT and ByteDance's Doubao in AI app usage.
What Is Qwen?
Qwen is a multimodal AI assistant built on Alibaba Cloud's Qwen model family. The consumer app handles text conversation, deep research, image generation, slide creation, and voice interaction. On the enterprise side, more than 90,000 businesses have deployed Qwen models through Alibaba Cloud Model Studio. The Qwen model family has surpassed 700 million total downloads on Hugging Face as of January 2026 — during December 2025, Qwen downloads exceeded the combined total of the next eight competing model families including Meta Llama, DeepSeek, and Mistral.
Qwen3.5, released in February 2026, now supports 201 languages (up from 82 in the previous generation), positions Qwen as one of the most multilingual AI platforms in existence. Benchmark performance is competitive with GPT-4o and Claude Sonnet on several reasoning and coding tasks, though independent evaluations vary by domain.
Who Makes Qwen?
Alibaba Cloud, the cloud computing division of Alibaba Group (NYSE: BABA / HKEX: 9988), headquartered in Hangzhou, China. The Qwen project has recently undergone significant leadership changes: technical lead Lin Junyang, post-training head Yu Bowen, and coding lead Hui Binyuan all departed in early 2026. Alibaba responded by centralising Qwen under top group leadership and forming a new task force to accelerate development. The commercial trajectory — 203 million MAUs, a unified brand identity — remains intact despite the management transition.
Key Features
- Qwen3.5 foundation model — Supports 201 languages, competitive with frontier models on reasoning and coding benchmarks, available as open-weight under Apache 2.0 license
- Deep research mode — Multi-step research agent that searches, synthesises, and presents findings with citations
- Image generation — AI image creation from text prompts integrated directly into the chat interface
- Slide creation — Automated presentation generation from text or document input
- Voice interaction — Real-time voice conversation using Qwen-Audio multimodal model
- Alibaba ecosystem integration — Connected to Taobao, Alipay, travel booking, and other Alibaba services for task execution beyond conversation
- 700M+ Hugging Face downloads — Qwen model family is the most downloaded open-weight model family on Hugging Face as of early 2026
Pricing
Source: chat.qwen.ai, verified March 2026.
- Free — Full access to Qwen consumer app including text, image generation, deep research, and voice
- Alibaba Cloud API — Qwen3.5 available via Alibaba Cloud Model Studio; Qwen-VL-Max priced at $0.41 per million input tokens. Enterprise volume pricing available.
- DingTalk Enterprise — Qwen integrated into Alibaba's enterprise workplace platform; over 2.2 million corporate users access Qwen through DingTalk
Qwen vs ChatGPT vs DeepSeek
Qwen3.5 is now a genuine frontier contender. On Hugging Face's open model rankings, Qwen3.5 variants occupied the top four positions globally in March 2026 — the first time any model family has held all four top spots simultaneously. Against ChatGPT, Qwen is free and more multilingual (201 languages vs GPT-4o's broader but less deep coverage), but OpenAI's model still leads on English-language reasoning and code generation in most benchmarks. Against DeepSeek, Qwen benefits from Alibaba's infrastructure and consumer distribution; DeepSeek leads on pure reasoning benchmark scores. Qwen's unique advantage is Alibaba ecosystem integration: the ability to execute tasks in e-commerce, payments, and services within the app is something neither ChatGPT nor DeepSeek can replicate.
Pros & Cons
Pros:
- 203 million MAUs in February 2026 — third largest AI app globally by usage
- Qwen3.5 is the most downloaded open-weight model family on Hugging Face
- 201-language support makes it the most multilingual major AI assistant available
- Free consumer app with no usage restrictions
- Alibaba ecosystem integration enables real-world task execution beyond chat
Cons:
- Senior leadership exodus in early 2026 (three key executives departed) creates execution uncertainty
- February 2026 MAU spike was partly promotional — organic retention at that scale is unproven
- English-language reasoning and coding lags GPT-4o and Claude Sonnet per independent benchmarks
- Data privacy considerations for non-Chinese users given Alibaba's regulatory environment
Who Should Use Qwen?
Qwen is the strongest choice for users who need multilingual AI — particularly across Asian languages where its training data depth exceeds Western competitors. Developers building on open-weight models should be running Qwen3.5 evaluations alongside Llama and Mistral; its download trajectory on Hugging Face signals genuine community adoption. For Chinese-market consumers and enterprises, Qwen's Alibaba ecosystem integration makes it more useful than ChatGPT for task completion in daily commerce. Western users should monitor the leadership transition before building production systems dependent on Alibaba Cloud's Qwen API.
Bottom Line
Qwen is the fastest-growing AI assistant of early 2026 and the most downloaded open-weight model family in the world. Its multilingual capability and Alibaba ecosystem integration are genuine competitive advantages. The leadership transition introduces uncertainty, but Alibaba's commercial commitment — 203 million users, a unified brand — means this is not a product in retreat. Watch Qwen3.5 benchmark comparisons closely; it is the first Chinese-developed model to genuinely compete with Western frontier models on a broad set of tasks.
