Qwen

Qwen is Alibaba's AI chat assistant and underlying model family, available as a conversational AI app on web and mobile. The consumer-facing product centers on a chat interface (currently running on a model the app itself labels "Qwen3.7-Plus" at time of review) where users can focus their questions and get direct answers. Who it's for: general users looking for a conversational AI assistant, with the official app specifically noting it's "designed for mobile devices, offering better experience and more features" — suggesting Alibaba is prioritizing the mobile experience as the primary way most people will use it. It also serves developers through the broader Qwen open-model family, widely used for self-hosting and fine-tuning outside the consumer chat app itself. Strengths: as one of the most prominent AI products out of China, Qwen benefits from Alibaba's infrastructure scale and represents one of the more actively developed non-US large-language-model efforts, with both a hosted chat product and an open-weight model family available for developers who want to run models themselves rather than depend on a single vendor's hosted service. Weaknesses and limitations: the consumer chat app's own web page is notably thin on self-description — it doesn't lay out specific capabilities, pricing, or a feature list on the login/landing screen itself, requiring account sign-up before most functionality becomes visible. This makes it harder to evaluate from the outside compared to competitors that publish a full features-and-pricing page before requiring signup. Real-world scenarios: using Qwen as a general chat assistant on mobile, and — for developers — downloading and self-hosting one of the open Qwen models (available through Hugging Face and similar hubs) to run inference on their own infrastructure rather than relying on a hosted API. Compared to ChatGPT, Claude, or Gemini, Qwen's main differentiators are its Alibaba-scale infrastructure and the availability of genuinely open model weights for self-hosting, which the more closed US competitors generally don't offer at the same scale. Behind the consumer chat app, the underlying Qwen3.5 model family (released February 2026) illustrates why Alibaba's models have gained so much traction among developers specifically: the flagship spans up to 397 billion parameters with support for 201 languages, and Qwen3-235B-A22B scores competitively with Gemini 2.5 Pro on benchmarks like ArenaHard and CodeForces, making it one of the strongest open-weight generalist models available. API pricing is genuinely aggressive too — the budget Qwen-Turbo tier runs $0.05/$0.20 per million input/output tokens, with the mid-tier Qwen-Plus at $0.40/$1.20 — though it's worth noting the fully free developer API tier was discontinued in April 2026, replaced by a one-time 70-million-token onboarding trial alongside the permanently free consumer chat app.